CoChat
Sign Up
Cited by 18Bluetooth and Wireless Communication TechnologiesWireless Body Area NetworksWireless Sensor Networks for Data Analysis

Comparison between Two Time Synchronization and Data Alignment Methods for Multi-Channel Wearable Biosensor Systems Using BLE Protocol

Open access
Wang, He, Li, Jianan, McDonald, Benjamin E., Farrell, Todd R., Huang, Xinming, Clancy, Edward A.·2023·Sensors
DOI 10.3390/s23052465PMC PMC10007376
Open this paper in CoChat — free accountOpen PDF

Abstract

Wireless wearable sensor systems for biomedical signal acquisition have developed rapidly in recent years. Multiple sensors are often deployed for monitoring common bioelectric signals, such as EEG (electroencephalogram), ECG (electrocardiogram), and EMG (electromyogram). Compared with ZigBee and low-power Wi-Fi, Bluetooth Low Energy (BLE) can be a more suitable wireless protocol for such systems. However, current time synchronization methods for BLE multi-channel systems, via either BLE beacon transmissions or additional hardware, cannot satisfy the requirements of high throughput with low latency, transferability between commercial devices, and low energy consumption. We developed a time synchronization and simple data alignment (SDA) algorithm, which was implemented in the BLE application layer without the need for additional hardware. We further developed a linear interpolation data alignment (LIDA) algorithm to improve upon SDA. We tested our algorithms using sinusoidal input signals at different frequencies (10 to 210 Hz in increments of 20 Hz—frequencies spanning much of the relevant range of EEG, ECG, and EMG signals) on Texas Instruments (TI) CC26XX family devices, with two peripheral nodes communicating with one central node. The analysis was performed offline. The lowest average (±standard deviation) absolute time alignment error between the two peripheral nodes achieved by the SDA algorithm was 384.3 ± 386.5 μs, while that of the LIDA algorithm was 189.9 ± 204.7 μs. For all sinusoidal frequencies tested, the performance of LIDA was always statistically better than that of SDA. These average alignment errors were quite low—well below one sample period for commonly acquired bioelectric signals.

AI snapshot

Written from the paper’s full text.

TL;DR

The paper compares simple data alignment (SDA) and linear interpolation data alignment (LIDA) for time-synchronizing multi-channel wearable biosensor nodes using BLE application-layer timestamps. Across 10–210 Hz sine-wave tests, LIDA consistently reduced absolute time alignment error relative to SDA, with both methods achieving sub-millisecond average errors.

Key findings

  • LIDA achieved a best average (±SD) absolute time alignment error of 189.9 ± 204.7 μs at 190 Hz, while SDA’s best was 384.3 ± 386.5 μs at 10 Hz.
  • Across all epochs, LIDA’s overall average absolute error was 0.38 ± 0.40 ms versus SDA’s 0.5442 ± 0.4481 ms; LIDA errors were lower than SDA in 24,756 of 38,115 epochs (65%), higher in 12,797 (34%), and tied in 562 (1%).
  • LIDA produced higher average correlation coefficients than SDA (0.9983 ± 0.0083 vs. 0.9945 ± 0.0261), and at every tested input frequency LIDA’s absolute time differences were significantly lower and correlation coefficients significantly higher (p = 10⁻⁵ or p < 10⁻⁵).
  • For both algorithms, 95th percentile errors were around 1–2 ms, typically about 1.5 ms, and average error did not trend with input frequency over 10–210 Hz.

Methods

  • The system used one TI CC2652R1 central node and two TI CC2640 peripheral nodes, each with a 12-bit ADC sampling at 1000 Hz; BLE application-layer timestamp pairs were generated and used with an affine regression model (N = 128 timestamp pairs, updated every K = 66 packets, 990 ms timestamp interval, 15 ms connection interval).
  • A signal generator produced common sine-wave inputs at 10–210 Hz in 20 Hz increments (11 frequencies), with 5 trials per frequency, 12 min per trial, and 55 total trials; data were transferred to a PC and analyzed offline in MATLAB.
  • After time synchronization and either SDA or LIDA alignment, the first 2 min were removed, remaining data were split into epochs of 100 sine cycles, upsampled 100×, and cross-correlation between the two peripheral channels was used to estimate absolute time alignment error.

Limitations

  • Stated: timestamp-pair precision can be impaired by blocked or delayed transmissions, and the authors did not archive timestamp pairs in this study, so they could not assess that characteristic in these data.
  • Stated: the study evaluated only two peripheral nodes, whereas EEG and high-density surface EMG systems may use more than 100 channels; LIDA’s computational cost on the central node may also limit use with more than two peripheral nodes.
  • Stated/evident: tests used a low-noise benchtop environment with no connection or packet loss and offline analysis of sine waves, so performance under real-world interference, packet loss, and real bioelectric signals was not directly tested.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3390/s23052465

Cites

39 of 39 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

10 of 34 cited sources checked · 3 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    BLE can be a more suitable wireless protocol than ZigBee and low-power Wi-Fi for multi-sensor wearable bioelectric signal acquisition systems.

    “Compared with ZigBee and low-power Wi-Fi, Bluetooth Low Energy (BLE) can be a more suitable wireless protocol for such systems.”
    • Also stated in: § 1. Introduction · plain
      “Compared with low-power Wi-Fi, BLE has lower power consumption of ~1 mA and a high transmission speed of 2 Mbps [17,18,19].”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites [17] Tosi, J., Taffoni, F., Santacatterina, M., Sannino, R., and Formica, D. (2017). Performance evaluation of bluetooth low energy: A systematic review. Sensors, 17. source

      The Introduction describes BLE as a candidate for medical body-area networks, including ECG and EMG applications. Section 3.3 reports a lower average current consumption than ZigBee, supporting one reason BLE might be suitable. The paper does not establish that BLE is more suitable than both ZigBee and low-power Wi-Fi specifically for multi-sensor wearable bioelectric signal acquisition. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites [18] Omre, 2010, Bluetooth low energy: Wireless connectivity for medical monitoring, J. Diabetes Sci. Technol. source

      The Introduction argues that BLE better meets medical-monitoring requirements than other available wireless technologies. It identifies ZigBee’s lower bandwidth and conventional Wi-Fi’s power consumption as drawbacks. It does not specifically assess low-power Wi-Fi or multi-sensor wearable bioelectric-signal acquisition systems, so the claim extends beyond the comparison made in the paper. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites [19] Fekr, 2015, Design and evaluation of an intelligent remote tidal volume variability monitoring system in e-health applications, IEEE J. Biomed. Health Inform. source
  • Premisecitation

    A prior method achieved high time synchronization precision of 39.92 ± 14.19 μs when a connection was established simultaneously for multiple nodes.

    “Alternatively, high time synchronization precision of 39.92 ± 14.19 µs was achieved when a connection was established simultaneously for multiple nodes [27].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [27] Bideaux, 2015, Synchronization in wireless biomedical-sensor networks with bluetooth low energy, Curr. Dir. Biomed. Eng. source
  • Premisecitation

    The frequency range of the input sine wave signals used in the performance tests (10-210 Hz) spans much of the relevant range of EEG, ECG, and EMG signals.

    “Hence, the frequency range of the input sine wave signals used in our performance tests spans much of the relevant range of these three common bioelectric signals.”
    Quote in the paperSubstantial claim2 of 6 sources checked
    • cites [33] Teplan, 2002, Fundamentals of EEG measurement, Measure. Sci. Rev.
    • cites [34] Bailey, 1990, Recommendations for standardization and specifications in automated electrocardiography: Bandwidth and digital signal processing, Circulation source
    • cites [35] Kligfield, 2007, Recommendations for the standardization and interpretation of the electrocardiogram: Part I: The electrocardiogram and its technology, Circulation source
    • cites [36] Hogan, 1980, Myoelectric signal processing: Optimal estimation applied to electromyography—Part II: Experimental demonstration of optimal myoprocessor performance, IEEE Trans. Biomed. Eng. source
    • cites [37] Merletti, 2020, Tutorial. Surface EMG detection, conditioning and pre-processing: Best practices, J. Electromyogr. Kinesiol. source

      The paper describes an sEMG bandwidth of 20–500 Hz (Analog low-pass, high-pass and notch filters; page unavailable), which overlaps the citing paper’s 10–210 Hz test range. Our copy does not establish the relevant frequency ranges for EEG or ECG, and 10–210 Hz does not cover the full stated sEMG bandwidth. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites [38] Gallina, 2022, Consensus for experimental design in electromyograpy (CEDE) project; high-density surface electromyography matrix, J. Electromyogr. Kinesiol. source

      The Sampling rate section gives an approximate 10–500 Hz bandwidth for high-density surface EMG, so the stated 10–210 Hz test range overlaps a substantial part of that range. Our copy does not establish the relevant frequency ranges for EEG or ECG, so it does not support the claim about all three signals. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
  • Findingown data

    The LIDA algorithm achieved a lowest average absolute time alignment error between two peripheral nodes of 189.9 ± 204.7 μs.

    “while that of the LIDA algorithm was 189.9 ± 204.7 μs.”
    • Also stated in: § 4.1. Overall Performance of SDA and LIDA Algorithms · plain
      “LIDA achieved its lowest average absolute time alignment error of 0.19 ± 0.20 ms at an input frequency of 190 Hz.”
    Quote in the paperSubstantial claim
  • Findingown data

    The SDA algorithm achieved a lowest average absolute time alignment error between two peripheral nodes of 384.3 ± 386.5 μs.

    “The lowest average (±standard deviation) absolute time alignment error between the two peripheral nodes achieved by the SDA algorithm was 384.3 ± 386.5 μs”
    • Also stated in: § 4.1. Overall Performance of SDA and LIDA Algorithms · plain
      “SDA achieved its lowest average time alignment error of 0.38 ± 0.39 ms at an input frequency of 10 Hz.”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim. 29
  • May combine two claims. 30
  • The quote only partly states this claim. 3

  • Classified as probably not a substantial claim.

    Premisecitation

    Many systems use beacon transmissions to achieve paired timestamps.

    “Many systems use beacon transmissions to achieve paired timestamps [26,27,30].”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites [26] Asgarian, F., and Najafi, K. (2017, January 22–24). Time synchronization in a network of bluetooth low energy beacons. Proceedings of the SIGCOMM Posters Demos, Los Angeles, CA, USA. source
    • cites [27] Bideaux, 2015, Synchronization in wireless biomedical-sensor networks with bluetooth low energy, Curr. Dir. Biomed. Eng. source

      The ‘Reference send receive delay’ section describes timestamps for transmitted and received data, and the Conclusion says that using the time difference between sending and receiving messages is common in synchronization protocols. Neither passage identifies those messages as beacon transmissions, so the cited paper supports the timestamping aspect but not the specified mechanism. (LLM, conf=0.77)

      Experimental — accuracy not yet measured
    • cites [30] Sridhar, S., Misra, P., and Warrior, J. (2015, January 13–16). Cheepsync: A time synchronization service for resource constrained bluetooth low energy advertisers. Proceedings of the 14th International Conference on Information Processing in Sensor Networks, Seattle, WA, USA. source

  • May combine two claims.

    Premisecitation

    BLE has lower power consumption (~1 mA) and a high transmission speed (2 Mbps) compared with low-power Wi-Fi.

    “Compared with low-power Wi-Fi, BLE has lower power consumption of ~1 mA and a high transmission speed of 2 Mbps [17,18,19].”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites [17] Tosi, J., Taffoni, F., Santacatterina, M., Sannino, R., and Formica, D. (2017). Performance evaluation of bluetooth low energy: A systematic review. Sensors, 17. source

      Section 2.1.1 identifies 2 Mbps as an optional BLE physical-layer data rate, not its typical application throughput. Section 3.3 describes BLE as low-power but reports current consumption that varies with operating conditions; it does not establish ~1 mA or a comparison specifically with low-power Wi-Fi. No page numbers are available in our copy. (LLM, conf=0.90)

      Experimental — accuracy not yet measured
    • cites [18] Omre, 2010, Bluetooth low energy: Wireless connectivity for medical monitoring, J. Diabetes Sci. Technol. source

      The Introduction describes BLE as low-power and Wi-Fi as relatively power-hungry in medical monitoring applications. But it gives BLE's bandwidth as 1 Mbps, not 2 Mbps, and its 1 mW radio output is not a ~1 mA current-consumption figure. It does not specifically compare BLE with low-power Wi-Fi. (LLM, conf=0.90)

      Experimental — accuracy not yet measured
    • cites [19] Fekr, 2015, Design and evaluation of an intelligent remote tidal volume variability monitoring system in e-health applications, IEEE J. Biomed. Health Inform. source

  • May combine two claims.

    Premisecitation

    Time synchronization can occur if multiple nodes are connected near-simultaneously, with time synchronization differences of 39.92 ± 14.19 μs.

    “Alternatively, time synchronization can occur if multiple nodes are connected near-simultaneously (time synchronization differences of 39.92 ± 14.19 µs [27]).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [27] Bideaux, 2015, Synchronization in wireless biomedical-sensor networks with bluetooth low energy, Curr. Dir. Biomed. Eng. source

  • May combine two claims. The quote only partly states this claim.

    Interpretationreasoning

    Longer latencies occur with more nodes, which can be detrimental to real-time control applications such as myoelectric prosthesis control or brain-computer interfaces.

    “Longer latencies occur, which can be detrimental to real-time control applications such as myoelectric prosthesis control [43] or brain–computer interfaces.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [43] Farrell, 2007, The optimal controller delay for myoelectric prostheses, IEEE Trans. Biomed. Eng. source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingcitation

    In the earlier study, most timestamp difference pairs clustered along the line of agreement, but a few noticeable outliers exist.

    “In fact, most do so, but a few noticeable outliers exist.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [29] Li, J. (2022). Advanced Control of Upper-Limb Prostheses with Time-Synchronized Distributed Wireless Electrodes. [Ph.D. Thesis, Worcester Polytechnic Institute].

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Findingnone given

    We have already transferred our SDA algorithm to run on the Nordic nRF5280 microcontroller unit.

    “In fact, we have already transferred our SDA algorithm to run on the Nordic nRF5280 microcontroller unit.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingcitation

    In the earlier study, timestamp pairs were available every 1050 ms during a 12 min trial, yielding 685 paired timestamps in total.

    “For that study, timestamp pairs were available every 1050 ms (this was the closest available interval to our 990 ms interval) during a 12 min trial (thus, there were 685 paired timestamps in total).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [29] Li, J. (2022). Advanced Control of Upper-Limb Prostheses with Time-Synchronized Distributed Wireless Electrodes. [Ph.D. Thesis, Worcester Polytechnic Institute].

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingnone given

    SDA has already been implemented on the central node of our real-time system and works reliably.

    “However, SDA has already been implemented on the central node of our real time system and works reliably.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    In preliminary testing, minimal data alignment performance differences were found as a function of the threshold values tested (1, 1.1, 1.25, 1.5, 1.75, 2, and 2.5 samples).

    “We found minimal data alignment performance differences as a function of the threshold values tested.”
    • Also stated in: § 2.3. Simple Data Alignment (SDA) Algorithm · plain
      “In preliminary testing (partial results available in [32]), we evaluated the threshold values (tc) of 1, 1.1, 1.25, 1.5, 1.75, 2, and 2.5 samples.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [32] Quintin, E., McDonald, B., Li, J., Wang, H., Clancy, T., and Farrell, T. (2021, January 9–11). Synchronization of multiple bluetooth low energy sensors for prosthetic control. Proceedings of the 2021 American Orthotic Prosthetic National Assembly, Boston, MA, USA.

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    A Kolmogorov-Smirnov test found the absolute time differences and the correlation coefficients each to be non-normally distributed (p < 10−5).

    “A Kolmogorov–Smirnov test found the absolute time differences and the correlation coefficients each to be non-normally distributed (p < 10⁻⁵).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Most distances between timestamp difference pairs and the line of agreement are under 0.5 ms, but large outliers exist.

    “Most distances are under 0.5 ms, but large outliers exist.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    All average correlation coefficients were very close to 1, with the lowest average correlation coefficient being 0.9919 when using SDA at an input frequency of 210 Hz.

    “However, all average correlation coefficients were very close to 1, with the lowest average correlation coefficient being 0.9919 when using SDA at an input frequency of 210 Hz.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    Our previous work found no strong dependence on the number of timestamp pairs, but an update interval of 750 ms was best.

    “We did not find strong dependence on number of timestamp pairs, but an update interval of 750 ms was best.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [29] Li, J. (2022). Advanced Control of Upper-Limb Prostheses with Time-Synchronized Distributed Wireless Electrodes. [Ph.D. Thesis, Worcester Polytechnic Institute].

      Not checked in the free run.

  • May combine two claims.

    Findingnone given

    The shapes of the distance histograms in Figure 5 are similar to those of the time synchronization histograms in Figure 4, but with smaller standard deviations.

    “Second, the shapes of the distance histograms in Figure 5 are similar to those of the time synchronization histograms in Figure 4, but with smaller standard deviations.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    Current time synchronization methods for BLE multi-channel systems via BLE beacon transmissions or additional hardware cannot satisfy the requirements of high throughput with low latency, transferability between commercial devices, and low energy consumption.

    “However, current time synchronization methods for BLE multi-channel systems, via either BLE beacon transmissions or additional hardware, cannot satisfy the requirements of high throughput with low latency, transferability between commercial devices, and low energy consumption.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    The additional hardware adds weight, volume, and battery consumption to the wearable device.

    “Additionally, the additional hardware adds weight, volume, and battery consumption to the wearable device.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationnone given

    Our method operates in the application layer, making it easily transferable to updated devices or those of other manufacturers.

    “In particular, our method operates in the application layer, making it easily transferable to updated devices or those of other manufacturers.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    For our test system consisting of only two peripheral nodes, the issue of primary clock loss was moot because we purposely maintained connection throughout each data collection trial.

    “For our test system consisting of only two peripheral nodes, this issue was moot, as we purposely maintained connection throughout each data collection trial.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Using the BLE beacon role, the central node periodically broadcasts clock information to peripheral nodes to maintain high synchronization accuracy (drift of <1 μs/min).

    “Using the beacon role, the central node periodically broadcasts clock information to the peripheral nodes to maintain high synchronization accuracy (drift of <1 µs/min [26]).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [26] Asgarian, F., and Najafi, K. (2017, January 22–24). Time synchronization in a network of bluetooth low energy beacons. Proceedings of the SIGCOMM Posters Demos, Los Angeles, CA, USA. source

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given

    The authors' time synchronization method is transferable between software versions and different manufacturers' devices.

    “Our method is transferable between software versions and different manufacturers’ devices.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Researchers have used custom add-on hardware to detect a synchronizing signal.

    “Alternatively, researchers have used custom add-on hardware to detect a synchronizing signal [28].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [28] Rheinländer, C.C., and Wehn, N. (November, January 30). Precise synchronization time stamp generation for bluetooth low energy. Proceedings of the 2016 IEEE Sensors, Orlando, FL, USA. source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    For the present work, in which about 2 min of timestamp data contributed to the affine time synchronization model, such large outliers could degrade time synchronization for up to about 4 min.

    “For the work described herein, in which ~2 min of timestamp data contributed to the affine time synchronization model, such large outliers could degrade time synchronization for up to ~4 min.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The average error did not trend as a function of input frequency, indicating that the implementation of both algorithms is feasible regardless of the frequency of the input signal over the range tested.

    “The average error did not trend as a function of input frequency, indicating that the implementation of both algorithms is feasible regardless of the frequency of the input signal (over the range of frequencies tested).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    An additional hardware approach provided time synchronization precision of 9 ± 17 μs.

    “One such approach provided time synchronization precision of 9 ± 17 µs [28].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [28] Rheinländer, C.C., and Wehn, N. (November, January 30). Precise synchronization time stamp generation for bluetooth low energy. Proceedings of the 2016 IEEE Sensors, Orlando, FL, USA. source

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    Customized hardware is not transferable for the large variety of commercial devices.

    “However, customized hardware is not transferable for the large variety of commercial devices.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationnone given

    Regulating the wireless environment in this manner is not possible in applied usage.

    “However, regulating the wireless environment in this manner is not possible in applied usage.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    The use of affine model time synchronization with a large timestamp buffer size smooths some possible deviations, but performance can still be impaired.

    “The use of affine model time synchronization with a large timestamp buffer size smooths some of these possible deviations, but the performance can still be impaired.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    High-precision data alignment relies entirely on the performance of time synchronization between peripheral nodes and the central node.

    “Of course, high-precision data alignment relies entirely on the performance of time synchronization between peripheral nodes and the central node.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationnone given

    The large outliers may be associated with rare blocked or delayed transmissions.

    “These large outliers may be associated with rare blocked or delayed transmissions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Study design and conductreasoning

    In the research reported herein, we set our connection interval at 15 ms, which was as short a connection interval as possible and facilitated reliable transmission using the TI CC26XX MCUs.

    “In the research reported herein, we set our connection interval at 15 ms, which was as short a connection interval as possible and facilitated reliable transmission using the TI CC26XX MCUs.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationcitation

    That prior method would either provide synchronization only once at startup, which would drift over time, or require continuously disconnecting and reconnecting devices, interrupting the data stream.

    “However, this method would either only provide synchronization once at startup (which would drift over time) or require continuously disconnecting and reconnecting devices, interrupting the data stream.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Previously published BLE time synchronization methods demonstrated lower errors than our methods but have clear disadvantages compared to our methods.

    “Previously published methods for BLE time synchronization (reviewed above) have demonstrated lower errors than our own, but have some clear disadvantages compared to our methods.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    We did not archive the timestamp pairs in this study, so could not assess whether successive central timestamp differences tightly match successive peripheral timestamp differences in our data.

    “Unfortunately, we did not archive the timestamp pairs in this study, so could not assess this characteristic in our data.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisenone given

    For the SDA algorithm, one primary peripheral clock needed to be selected to which all other peripheral clocks were aligned.

    “For our SDA algorithm, we needed to select one primary peripheral clock to which all other peripheral clocks were aligned.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Each peripheral node has its own ADC following its local clock, and each ADC runs at a slightly different sampling rate and phase.

    “Each peripheral node consists of its own analog-to-digital-converter (ADC) which follows its local clock, and each ADC will run at a slightly different sampling rate and phase [21].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [21] Araujo, 2019, Performance of clock sources and their influence on time synchronization in wireless sensor networks, Int. J. Distrib. Sens. Netw.

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The SDA algorithm achieved precise time synchronization and data alignment (error of approximately one sample period) in real time.

    “We found that a simple data alignment algorithm (SDA) achieved precise time synchronization and data alignment (error of approximately one sample period) in real time [29].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [29] Li, J. (2022). Advanced Control of Upper-Limb Prostheses with Time-Synchronized Distributed Wireless Electrodes. [Ph.D. Thesis, Worcester Polytechnic Institute].

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    When the central node broadcasts a beacon transmission, it cannot receive data from peripheral nodes, risking interruption of high-bandwidth continuous peripheral data streams and data loss.

    “However, when the central node broadcasts a beacon transmission, it cannot receive data from peripheral nodes, thus risking interruption of the high-bandwidth, continuous peripheral data streams needed for these bioelectric signals. Data loss could occur.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    Our method operates in the application layer, so it is easier to program and is transferable between devices and software versions.

    “Our method reported herein operates in the application layer, so is easier to program and is transferable between devices and software versions.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationnone given

    Another clear advantage of LIDA is that data from each peripheral node are independently aligned directly to time on the central node, and data alignment from one peripheral node is not a function of any other peripheral node.

    “As noted above, another clear advantage of the LIDA algorithm is that data from each peripheral node are independently aligned directly to time on the central node. Data alignment from one peripheral node is not a function of any other peripheral node.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    In our tests, we consistently maintained two peripheral nodes connected to one central node without connection or packet loss.

    “In our tests, we consistently maintained two peripheral nodes connected to one central node without connection or packet loss.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    The least squares method of parameter optimization minimizes mean squared estimation error, is non-iterative with no expected convergence issues, and has low computational cost.

    “We chose the least squares method of parameter optimization because it minimizes the mean squared estimation error, is a non-iterative optimization technique with no expected convergence issues (and is thus robust), and has a low computational cost.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning

    With a connection interval of 15 ms, a packet has been dropped if the ADC timestamp difference between successive packets from a given peripheral node is 30 ms, and two packets have been dropped if the difference is 45 ms.

    “For example, with our connection interval of 15 ms, a packet has been dropped if the ADC timestamp difference between successive packets from a given peripheral node is 30 ms. Two packets have been dropped if the ADC timestamp difference is 45 ms, etc.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Timestamp errors from the two peripherals would likely have an additive influence on time synchronization, leading to larger standard deviations, which is consistent with our results.

    “However, the timestamp errors from the two peripherals would likely have an additive influence on time synchronization (as each peripheral node timestamp error grows, so does the time synchronization error). Additive error effects lead to larger standard deviations; this is consistent with our results.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    We eliminated all other active BLE devices from the environment to the extent possible.

    “In fact, to the extent possible, we eliminated all other active BLE devices from the environment.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    Custom design is required and is specific to the implemented system, so the system may require a complete re-design as each new hardware microcontroller or BLE software version is released.

    “Custom design is required and is specific to that implemented system. Hence, the system may require a complete re-design as each new hardware microcontroller or BLE software version is released (which occurs quite frequently, i.e., yearly, or even more frequently in some cases).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationnone given

    Additional hardware increases power consumption, which decreases the battery life of the wireless wearable system in daily use.

    “The additional hardware also increases power consumption, which decreases the battery life of the wireless wearable system in daily use.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    We only evaluated time synchronization performance using two peripheral nodes.

    “Another limitation of this research is that we only evaluated time synchronization performance using two peripheral nodes.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Each method provided average data alignment errors ≤0.70 ms, which is sufficient for many target applications, including many EEG, ECG, and EMG acquisition systems.

    “Nevertheless, each method provided average (±standard deviation) data alignment errors ≤0.70 ms, which is sufficient for many target applications, including many EEG, ECG, and EMG acquisition systems.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    If too many channels are wired to the same peripheral, the number of samples per connection interval may exceed the packet size.

    “If too many channels are wired to the same peripheral, the number of samples per connection interval may exceed the packet size.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    As more peripheral nodes are added, the connection interval will need to grow, at least linearly and perhaps at a steeper pace due to increased packet size, since the central node must receive data from each peripheral node serially.

    “As more peripheral nodes are added, the connection interval will need to grow (at least linearly, perhaps at a steeper pace due to the increased packet size), since the central node must receive data from each peripheral node in a serial fashion (i.e., time-division multiplexed).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). The quote only partly states this claim.

    Interpretationnone given

    Our general approach to time synchronization and data alignment should remain useful for multichannel systems.

    “In any case, our general approach to time synchronization and data alignment should remain useful.”
    Quote in the paperNot a substantial claim
51 of 51

  • The paper relies on this without support.

    ZigBee does not meet the bandwidth requirement for bioelectric signals such as EEG, ECG, and EMG, especially EMG.

    “ZigBee does not meet the bandwidth requirement for these bioelectric signals, especially for EMG signals.”

  • The paper relies on this without support.

    The timestamp pair TSc and TSp represent the same actual time (or times with a consistent offset) on the central and peripheral devices.

    “The timestamp pair TSc and TSp, to the best of their ability, represent the same times on each of the central and peripheral devices.”

  • The paper relies on this without support.

    The central timestamp incremented by one transmit interval is an excellent estimate of the central timestamp corresponding to the ensuing BLE notification when received on that specific peripheral node.

    “This timestamp is incremented by one transmit interval and forms an excellent estimate of the central timestamp (TSc) corresponding to the ensuing BLE notification when received on that specific peripheral node.”

  • The paper relies on this without support.

    The ADC timestamp closely corresponds to the peripheral time associated with the last ADC sample in the packet.

    “This timestamp (TSADC) closely corresponds to the peripheral time associated with the last ADC sample in the packet”

  • The paper relies on this without support.

    The affine regression model accurately estimates central time from peripheral time.

    “The N most recent timestamp pairs from each peripheral node were used to synchronize their respective times to that of the central clock via an affine regression model that estimated central time from peripheral time.”

  • The paper relies on this without support.

    The slope parameters β1,m will have values very close to 1.0.

    “Note that slope parameters β1,m will have values very close to 1.0 (depending on the accuracies of the central and peripheral node clocks).”

  • The paper relies on this without support.

    The cross-correlation maximum location estimates the absolute time alignment error between data streams from the two peripheral nodes.

    “For each epoch, the absolute value, in ms, of the location of the correlation maximum was the estimate of the absolute time alignment error between data streams from the two peripheral nodes.”

  • The paper relies on this without support.

    BLE message exchange times are not tightly controlled in the small MCUs used in wearable devices.

    “However, BLE message exchange times are not tightly controlled in the small MCUs used in wearable devices, making this technique inappropriate.”

  • The paper relies on this without support.

    New MCUs are being offered frequently, often in timeframes of less than one year, by an array of manufacturers, with newer model performance leapfrogging that of existing models.

    “New MCUs are being offered frequently, often in timeframes of less than one year, by an array of manufacturers, with newer model performance “leapfrogging” that of existing models.”

  • The paper relies on this without support.

    The more wireless nodes operate in a system, the higher the likelihood of delayed or blocked transmissions, which can lead to degradation in the quality of the paired timestamps.

    “Of course, the more wireless nodes operate in a system, the higher the likelihood of delayed or blocked transmissions—which can lead to degradation in the quality of the paired timestamps.”

  • The paper relies on this without support.

    Many bioelectric systems, especially EEG and high-density surface EMG, can include more than 100 recording channels.

    “Many bioelectric systems (especially EEG and high-density surface EMG) can include more than 100 recording channels.”
11 of 11

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Figure 3 is cited for a comparison it does not show

    Minor

    The paper describes Figure 3 as showing unaligned signals and signals after LIDA, not signals aligned with SDA. It therefore cannot show that LIDA errors are closer to zero than SDA errors.

1 of 1

Cites

Flagged first, then by citations

Research works printed but not linked: 4

  1. Li, J. (2022). Advanced Control of Upper-Limb Prostheses with Time-Synchronized Distributed Wireless Electrodes. [Ph.D. Thesis, Worcester Polytechnic Institute].
  2. Quintin, E., McDonald, B., Li, J., Wang, H., Clancy, T., and Farrell, T. (2021, January 9–11). Synchronization of multiple bluetooth low energy sensors for prosthetic control. Proceedings of the 2021 American Orthotic Prosthetic National Assembly, Boston, MA, USA.
  3. Proakis, J.G., and Manolakis, D.G. (2022). Digital Signal Processing: Principles, Algorithms, and Applications, Pearson Education, Inc.
  4. Ballantine, 1952, Distance from a line, or plane, to a point, Am. Math. Mon., 59, 242

Cited by

Still being checked.

Related

Authors
Wang, H., Li, J., McDonald, B. E., Farrell, T. R., Huang, X., & Clancy, E. A.
Affiliations
  • Worcester Polytechnic Institute (US)
  • Liberating Technologies (United States) (US)
Corresponding author
Clancy, Edward A.
ORCID
Venue
Sensors
Publisher
MDPI AG
Published
2023-02-23
Volume
23
Issue
5
Pages
2465
ISSN
1424-8220 (ISSN-L)
DOI
10.3390/s23052465
PMC
PMC10007376
OpenAlex
W4321597723
Semantic Scholar
ecec759baa7f514c2c29e1c41bcd8c6c90c26eb8
Cited by
18
References
39
Citations by year
2026: 9 · 2025: 6 · 2024: 1 · 2023: 2
FWCI
2.00
Type
journal article
Language
en
Open access
Licence
Locations
  • Sensors (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
US Army Medical Research and Materiel Command (W81XWH-18-C-0111)
Topics
  • Physical Sciences › Computer Science › Computer Networks and Communications › Bluetooth and Wireless Communication Technologies
  • Physical Sciences › Engineering › Biomedical Engineering › Wireless Body Area Networks
  • Physical Sciences › Computer Science › Computer Networks and Communications › Wireless Sensor Networks for Data Analysis
Keywords
Bluetooth Low Energy, clock synchronization, data alignment, bioelectric signals, EEG, ECG, electromyography, wireless sensor networks, low latency
MeSH terms
Wearable Electronic Devices, Acetanilides, Acetates, Algorithms, Biosensing Techniques, Wireless Technology

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-28.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-28.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review