Validation of Player and Ball Tracking with a Local Positioning System

Predicting Field-Sport Distances Without Global Positioning Systems in Indoor Play: A Comparative Study of Machine-Learning Techniques

PURPOSE

Accurately predicting the distance covered by athletes during indoor sport activities without the use of GPS (global positioning systems) presents a significant challenge. This study evaluates the effectiveness of various machine-learning techniques in predicting total distance, sprinting distance, and running distance for athletes in men's and women's soccer and lacrosse at the University of Notre Dame.

METHODS

The techniques assessed include XGBoost Regressor, ElasticNet Regression, Ridge Regression, and Lasso Regression. Key performance metrics such as root-mean-square error, SDs, means, and 95% CIs are analyzed to provide insights into the relative performance of each method.

RESULTS

XGBoost provided the lowest root-mean-square error for total distance (97.962 [12.973]), sprint distance (91.616 [4.234]), and running distance (137.103 [2.789]).

CONCLUSION

The results demonstrate varying levels of accuracy and precision across different sports, genders, and contexts (game vs practice), highlighting the importance of selecting the appropriate model for specific applications in optimizing team performance, injury prevention, and player conditioning.

Tracking Methods in Sports: a Review of Advances, Quality, and Challenges in Performance Data

Tracking systems in sports aim to record the athlete's position as a function of time. From these data, information on physical, tactical and technical performance is obtained and assists coaches and players in decision-making during the training and competition routine. The implementation, feasibility, and quality of data generated by tracking systems depend on the conditions of each sporting environment and their requirements. This narrative review addresses the fundamentals of the main tracking systems, including algorithms based on computer vision and artificial intelligence for processing videos and global (global positioning system and global navigation satellite system) and local positioning systems. We also address technological advances for obtaining data from human pose estimation and the main validation or quality analysis studies of each method. Finally, we present a series of recommendations and future directions for the evaluation and development of automatic and accurate athlete tracking tools.

Agility in Handball: Position- and Age-Specific Insights in Performance and Kinematics Using Proximity and Wearable Inertial Sensors

Handball is a dynamic team sport characterized by high agility requirements, which feature complex motor-cognitive demands. The ability to meet these demands is critical for performance in handball but remains under-represented in research. Existing studies highlight that cognitive demands can strongly interfere with motor behavior, particularly in dynamic sport-specific movement tasks. Furthermore, high motor-cognitive load is associated with risk of lower limb injury. Therefore, to gain insight in the mechanisms between movement and performance dynamics in the presence of cognitive demands, this study investigated the performance of elite handball players in a novel planned and reactive agility task. Four FitLight proximity sensors (FitLight Corp, Aurora, ON, Canada) recorded execution time. Nine Noraxon Myomotion wearable inertial sensors (Noraxon U.S.A. Inc., Scottsdale, AZ, USA) tracked the motion of the players' trunk, pelvis, and lower extremities at 200 Hz. Execution time and kinematics were compared between adult and youth players and between different playing positions. Adult players demonstrated faster performance than youth players and exhibited differences in hip and knee flexion, potentially reflecting variations in acceleration and deceleration strategies. Backcourt players and wings demonstrated faster performance compared to pivots, who showed distinct patterns of hip, knee, and ankle flexion, possibly due to differences in body composition. These findings highlight the influence of motor and cognitive demands on agility performance and offer valuable insights into age- and position-specific differences among elite handball players. Furthermore, these findings support the use of wearable inertial sensors for the purpose of athlete evaluation. Future research should explore the implementation into athlete monitoring.

Validation of the metabolic power model during three intermittent running-based exercises with emphasis on aerobic and anaerobic energy supply

Introduction

In intermittent sports, available internal load measurements like capillary blood techniques and portable respiratory gas analyzers are considered as gold standards in controlled laboratory environments, but are impractical for daily use in training and matches. A newer approach, the metabolic power model, allows to extrapolate from speed and acceleration data to the metabolic power, simulated oxygen uptake, and aerobic and anaerobic energy supply. The aim of this study was to validate the metabolic power model against the established 3-component model to allow direct comparison of variables including energy expenditure and supplies during intermittent running-based exercises.

Methods

Twelve male athletes (24 ± 3 years) performed three different running-based exercises consisting of continuous shuttle runs and repeated accelerations and sprints with change of direction. Each exercise condition intended to primarily stress the aerobic, anaerobic alactic, and lactic energy supply. One-way repeated measures ANOVA or Friedman test and corresponding effect sizes were applied for statistical analyses. Additionally, absolute and relative biases and Bland-Altman plots were generated.

Results

For total energy expenditure, there were statistically significant differences (p ≤ .002, d ≥ .882, large) and biases of -13.5 ± 11.8% for the continuous shuttle runs and up to 352.2 ± 115.9% for repeated accelerations and sprints. Concerning aerobic energy supply, there were statistically significant differences (p < .001, d ≥ 1.937, large effect sizes) and biases of up to -38.1 ± 11.7%. For anaerobic energy supply, there were statistically significant differences (p < .001, d ≥ 5.465, large) and biases of up to 1,849.9 ± 831.8%.

Discussion

In conclusion, the metabolic power model significantly under- or overestimates total energy expenditure and supplies with large effect sizes during intermittent running-based exercises. Future studies should optimize the model before it can be used on a daily basis for scientific and practical purposes.

An integrated dataset of spatiotemporal and event data in elite soccer

Data-driven match analysis in soccer is a growing discipline in both research and practice. However, public data is scarce, which raises the barrier for entering this field and decreases reproducibility of methods and results. To bridge this gap, this paper presents a dataset of official match information, event, and position data from seven matches of the German Bundesliga's first and second division. The match information contains meta data about the matches and their participants. The event data contain timestamps along with descriptions of discrete events, like passes, shots, or fouls. The position data contain the x/y-coordinates of every player and the ball. By integrating multiple data modalities - i.e., event logs with timestamps, and x-y coordinates of player and ball positions - the dataset offers a multidimensional view of match dynamics. This dataset supports the validation of existing analytical techniques and facilitates the development of new methodologies in sports analytics. With availability under CC-BY 4.0, it promotes transparency, reproducibility, and the idea of open science in match analysis research.

Accuracy Standards of Wearable Technologies for Assessment of Soccer Kicking: Protocol for a Systematic Literature Review

BACKGROUND

Wearable technology is widely applied in performance monitoring, an integral part of sports and exercise sciences. The kick movement in soccer exemplifies a sports technique that could benefit from appropriate biomechanics assessment methodologies. However, the accuracy of wearables in quantifying soccer kick mechanics, particularly under field conditions, remains unclear.

OBJECTIVE

This study aims to present a protocol for a systematic review to discuss the measurement properties (validity, reliability, and/or accuracy aspects) of wearable technology systems explicitly used to measure ball-kicking features in soccer.

METHODS

This review protocol was preregistered in the Open Science Framework. A total of 2 authors will perform searches in major electronic databases using specific keyword combinations in PubMed, Physical Therapy and Sports Medicine, Web of Science, ProQuest, IEEE Xplore, EBSCOHost, and Scopus. Following a specific population, intervention, comparison, outcome framework (population: soccer players and/or collected human data in a football-related environment; intervention: at least 1 wearable used; comparator: criterion measures, repeated testing sessions and/or actual values; outcome: ball kicking data), studies will be screened based on predetermined inclusion and exclusion criteria. The methodological quality of the included studies will be assessed using the "consensus-based standards for the selection of health measurement instruments" checklist (in studies concerning validity or reliability) or the "quality assessment of diagnostic accuracy studies" tool (in studies concerning accuracy). Data extraction will be conducted to determine the level of evidence according to the "best evidence synthesis method," and an evidence gap map will be constructed. The Cohen κ coefficient will be used to estimate the interevaluator agreement.

RESULTS

This ongoing systematic review has completed database searches and is currently in the screening phase. Depending on the number and consistency of studies, results may be presented by meta-analysis or qualitative synthesis, with subgroup analyses considering factors such as gender, age, and playing level. The final results are expected by July 2024, with manuscript submission anticipated by November 2024.

CONCLUSIONS

Our study will provide a comprehensive summary of the highest level of evidence available on the use of wearables for the assessment of soccer kick mechanics, providing practical guidance for athletes and sports sciences professionals regarding the validity and reliability aspects of using wearable technology to measure ball-kicking features in soccer.

TRIAL REGISTRATION

OSF registries https://osf.io/zm3j6.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

DERR1-10.2196/57433.

Concurrent validity and between-unit reliability of a foot-mounted inertial measurement unit to measure velocity during team sport activity

The concurrent validity and between-unit reliability of a foot-mounted inertial measurement unit (F-IMU) was investigated during linear and change of direction running drills. Sixteen individuals performed four repetitions of two drills (maximal acceleration and flying 10 m sprint) and five repetitions of a multi-directional movement protocol. Participants wore two F-IMUs (Playermaker) and 10 retro-reflective markers to allow for comparisons to the criterion system (Qualisys). Validity of the F-IMU derived velocity was assessed via root-mean-square error (RMSE), 95% limits of agreement (LoA) and mean difference with 95% confidence interval (CI). Between-unit reliability was assessed via intraclass correlation (ICC) with 90% CI and 95% LoA. The mean difference for instantaneous velocity for all participants and drills combined was -0.048 ± 0.581 m ∙ s-1, the LoA were from -1.09 to -1.186 m ∙ s-1 and RMSE was 0.583 m ∙ s-1. The ICC ranged from 0.84 to 1, with LoA from -7.412 to 2.924 m ∙ s-1. Differences were dependent on the reference speed, with the greatest absolute difference (-0.66 m ∙ s-1) found at velocities above 7 m ∙ s-1. Between-unit reliability of the F-IMU ranges from good to excellent for all locomotor characteristics. Playermaker has good agreement with 3D motion capture for velocity and good to excellent between-unit reliability.

Agreement and Sensitivity of the Acceleration-Velocity Profile Derived via Local Positioning System

Sprint performance is commonly assessed via discrete sprint tests and analyzed through kinematic estimates modeled using a mono-exponential equation, including estimated maximal sprinting speed (MSS), relative acceleration (TAU), maximum acceleration (MAC), and relative propulsive maximal power (PMAX). The acceleration-velocity profile (AVP) provides a simple summary of short sprint performance using two parameters: MSS and MAC, which are useful for simplifying descriptions of sprint performance, comparison between athletes and groups of athletes, and estimating changes in performance over time or due to training intervention. However, discrete testing poses logistical challenges and defines an athlete's AVP exclusively from the performance achieved in an isolated testing environment. Recently, an in situ AVP (velocity-acceleration method) was proposed to estimate kinematic parameters from velocity and acceleration data obtained via global or local positioning systems (GPS/LPS) over multiple training sessions, plausibly improving the time efficiency of sprint monitoring and increasing the sample size that defines the athlete's AVP. However, the validity and sensitivity of estimates derived from the velocity-acceleration method in relation to changes in criterion scores remain elusive. To assess the concurrent validity and sensitivity of kinematic measures from the velocity-acceleration method, 31 elite youth basketball athletes (23 males and 8 females) completed two maximal effort 30 m sprint trials. Performance was simultaneously measured by a laser gun and an LPS (Kinexon), with kinematic parameters estimated using the time-velocity and velocity-acceleration methods. Agreement (%Bias) between laser gun and LPS-derived estimates was within the practically significant magnitude (±5%), while confidence intervals for the percentage mean absolute difference (%MAD) overlapped practical significance for TAU, MAC, and PMAX using the velocity-acceleration method. Only the MSS parameter showed a sensitivity (%MDC95) within practical significance (<5%), with all other parameters showing unsatisfactory sensitivity (>10%) for both the time-velocity and velocity-acceleration methods. Thus, sports practitioners may be confident in the concurrent validity and sensitivity of MSS estimates derived in situ using the velocity-acceleration method, while caution should be applied when using this method to infer an athlete's maximal acceleration capabilities.

The burdens of sitting on the bench - comparison of absolute and relative match physical load between handball players with high and low court time and implications for compensatory training

Wearables quantify the activity in team sports and indicate that players experience peak physical loads during competitions. Accordingly, players with limited court time in competitions will miss important training stimuli. The present study aimed to quantify these gaps in physical load in professional handball players. Activity of all players competing in the 2021/2022 Bundesliga (Germany) was tracked using Kinexon LPS sensors. Gaps in physical load were quantified comparing the 25% of appearances with the highest (HIGH; 51.8 ± 5.2 mins) and lowest court times (LOW; 10.1 ± 4.3 mins). Distances, accumulated acceleration, jumps, sprints, impacts, accelerations, and decelerations were analysed as absolute and relative (per minute) outcomes. Players were grouped into wings, backcourts, and pivots. Unpaired t-tests between HIGH and LOW were performed (p < .05), and effect sizes were calculated (Cohen´s d). Analyses revealed significant effects of court time on activity. While absolute activity increased for HIGH, relative activity increased for LOW (p < .05). In addition, effect sizes revealed position-specific gaps in physical load, particularly for acyclic activities (jumps, accelerations). Gaps in physical load resulting from limited court time are highly position-specific. Our observations may provide benchmarks for the position-specific calibration of compensatory training.

Analysis of contextualized intensity in Men's elite handball using graph-based deep learning

Manual annotation of data in invasion games is a costly task which poses a natural limit on sample sizes and the level of granularity used in match and performance analyses. To overcome this challenge, this work introduces FAUPA-ML, a Framework for Automatic Upscaled Performance Analysis with Machine Learning, which leverages graph neural networks to scale domain-specific expert knowledge to large data sets. Networks were trained using position data of match phases (counter/position attacks), annotated manually by domain experts in 10 matches. The best network was applied to contextualize N = 539 matches of elite handball (2019/20-2021/22 German Men's Handball Bundesliga) with 86% balanced accuracy. Distance covered, speed, metabolic power, and metabolic work were calculated for attackers and defenders and differences between counters and position attacks across seasons analyzed with an ANOVA. Results showed that counter attacks are shorter, less frequent and more intense than position attacks and that attacking is more intense than defending. Findings show that FAUPA-ML generates accurate replications of expert knowledge that can be used to gain insights in performance analysis previously deemed infeasible. Future studies can use FAUPA-ML for large-scale, contextualized analyses that investigate influences of team strength, score-line, or team tactics on performance.

Analysis of Motion Characteristics and Metabolic Power in Elite Male Handball Players

While handball is characterized by repeated sprints and changes of direction, traditional player load models do not consider accelerations and decelerations. The aim of this study was to analyze the differences between metabolic power and speed zones for player load assessment with regard to the player role. Position data from 330 male individuals during 77 games from the 2019/20 German Men's Handball-Bundesliga (HBL) were analyzed, resulting in 2233 individual observations. Players were categorized into wings, backs and pivots. Distance covered in different speed zones, metabolic power, metabolic work, equivalent distance (metabolic work divided by energy cost of running), time spend running, energy spend running, and time over 10 and 20 W were calculated. A 2-by-3 mixed ANOVA was calculated to investigate differences and interactions between groups and player load models. Results showed that total distance was longest in wings (3568 ± 1459 m in 42 ± 17 min), followed by backs (2462 ± 1145 m in 29 ± 14 min), and pivots (2445 ± 1052 m in 30 ± 13 min). Equivalent distance was greatest in wings (4072.50 ± 1644.83 m), followed by backs (2765.23 ± 1252.44 m), and pivots (2697.98 ± 1153.16 m). Distance covered and equivalent distance showed moderate to large interaction effects between wings and backs (p < .01, ES = 0.73) and between wings and pivots (p < .01, ES = 0.86) and a small interaction effect between backs and pivots (p < .01, ES = 0.22). The results underline the need for individualized management of training loads and the potential of using information about locomotive accelerations and decelerations to obtain more precise descriptions of player load during handball game performance at the highest level of competition. Future studies should investigate the influence of physical performance on smaller match sequences, like ball possession phases.

Physical match demands of four LIQUI-MOLY Handball-Bundesliga teams from 2019-2022: effects of season, team, match outcome, playing position, and halftime

Introduction

Due to the development in team handball, there is a need to optimize the physical capacities of team handball players for which knowledge of the physical match demands is essential. The aim of this study was to investigate the physical match demands of four LIQUI-MOLY Handball-Bundesliga (HBL) teams across three seasons with respect to the effects of season, team, match outcome, playing position, and halftime.

Methods

A fixed installed local positioning system (Kinexon) was used, collecting 2D positional and 3D inertial measurement unit data at 20 and 100 Hz, respectively. The physical match demands were operationalized by basic (e.g., distance, speed, and acceleration) and more advanced variables (e.g., jumps, throws, impacts, acceleration load, and metabolic power). A total of 347 matches (213 with an additional ball tracking) were analyzed from four teams (one top, two middle, and one lower ranked) during three consecutive seasons (2019-2022). One-way ANOVAs were calculated to estimate differences between more than two groups (e.g., season, team, match outcome, playing position). Mean differences between halftimes were estimated using Yuen's test for paired samples.

Results

Large effects were detected for the season (0.6 ≤ ξ ^ ≤ 0.86 ), team (0.56 ≤ ξ ^ ≤ 0.72 ), and playing position (0.64 ≤ ξ ^ ≤ 0.98 ). Medium effects were found for match outcome (ξ ^ ≤ 0.36 ) and halftime (ξ ^ ≤ 0.47 ).

Conclusion

For the first time, we provide a comprehensive analysis of physical match demands in handball players competing in the LIQUI-MOLY Handball-Bundesliga. We found that physical match demands differ on that top-level with up to large effect sizes concerning the season, team, match outcome, playing position, and halftime. Our outcomes can help practitioners and researchers to develop team and player profiles as well as to optimize talent identification, training, regeneration, prevention, and rehabilitation procedures.

Accuracy of a local positioning system for time-series speed and acceleration and performance indicators in game sports

The objective was to determine the reliability and validity of a local positioning system (LPS) promising high accuracy at reduced product costs. Fifty-five random static positions in a gym (54.8  × 26.0 m) were obtained 10 times via LPS (50 Hz) and measuring tape. An athlete's LPS-derived peak and time-series speed and acceleration during dynamic movements (n = 80) were compared with Vicon (100 Hz). Reliability and validity were assessed via Intraclass and Concordance Correlation Coefficients (ICC/CCC), root mean square errors, Bland-Altman plots, and analysis of variance. ICC3,1 (≥0.999) and CCC (0.387-0.999) were calculated for static positions (errors <0.22 m). CCC for time-series speed and acceleration, and peak speed, acceleration, and deceleration were 0.884-0.902, 0.777-0.854, 0.923, 0.486, and 0.731, respectively. Errors were larger in time-series acceleration (14.37 ± 3.77%) than in speed (11.99 ± 5.78%) (η p 2  = 0.472, p < 0.001) and in peak acceleration (28.04 ± 14.34%) and deceleration (25.07 ± 14.90%) than in speed (7.34 ± 6.07%) (η p 2  = 0.091, p < 0.01). LPS achieved excellent reliability and moderate-to-excellent validity of time-series speed and acceleration. The system accurately measured peak speed but not peak acceleration and deceleration. The system is suitable for analyses based on instantaneous speed and acceleration in game sports (e.g., energy estimations).

Within-week differences in external training load demands in elite volleyball players

Purpose

The aim of this study was to analyze the within-week differences in external training intensity in different microcycles considering different playing positions in women elite volleyball players.

Methods

The training and match data were collected during the 2020–2021 season, which included 10 friendly matches, 41 league matches and 11 champions league matches. The players’ position, training/match duration, training/match load, local positioning system (LPS) total distance, LPS jumps, accelerations, decelerations, high metabolic load distance (HMLD), acute and chronic (AC) mean and AC ratio calculated with the rolling average (RA) method and the exponentially weighted moving average (EWMA) method, monotony and strain values were analyzed.

Results

All the variables except strain, Acc/Dec ratio and acute mean (RA) showed significant differences among distance to match days. Regarding the players’ positions, the only difference was found in the AC ratio (EWMA); in all microcycles, the middle blocker player showed workload values when compared with the left hitter, setter and libero.

Conclusion

Overall, the analysis revealed that the intensity of all performance indicators, except for strain, acc/dec and acute mean load (RA), showed significant differences among distance to match day with moderate to large effect sizes. When comparing players’ positions, the middle blocker accumulated the lowest loads. There were no significant differences among other positions.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13102-022-00568-1.

Local Positioning System-Derived External Load of Female and Male Varsity Ice Hockey Players During Regular Season Games

Purpose

The purposes of this study were to quantify the external load for female and male varsity ice hockey players during regular season games using a local positioning system (LPS), compare LPS-derived external load between sexes and positions, and compare skating distances in absolute and relative speed zones.

Methods

Data were collected for 21 female (7 defense, 14 forwards; 20.0 ± 1.4 yrs., 69.1 ± 6.7 kg, 167.1 ± 5.4 cm) and 25 male (8 defense, 17 forwards; 21.9 ± 1.1 yrs., 85.9 ± 5.4 kg, 181.1 ± 5.2 cm) varsity ice hockey players. Measures included skating distance (total, and in absolute and relative speed zones), peak skating speed, peak acceleration and deceleration, accumulative acceleration load, and number of accelerations, decelerations, turns, skating transitions, direction changes, and impacts.

Results

Female and male players had a high external load during games, with average peak skating speeds >28 km/h and average skating distances >4.4 km. Most LPS-derived measures showed greater external load in males than females (p < 0.05). Forwards skated further at higher speeds compared to defense in both sexes (p < 0.001). Skating distances were significantly different when comparing absolute and relative speed zones (p < 0.001), with absolute speed zones potentially overestimating skating at very slow, very fast, and sprint speeds and underestimating skating at slow and moderate speeds.

Conclusion

This was the first study to measure external load in female ice hockey players with a LPS. Both female and male varsity players had high external loads during games, with forwards having greater external load at higher intensities and defense having greater external load at lower intensities. Sex and positional differences outline the importance of individualized athlete monitoring.

Throwing performance by playing positions of male handball players during the European Championship 2020

The aim of this study was to analyze the throwing velocity and effectiveness of elite male handball players during the European Championship 2020. A local positioning system was used for the first time in a European Championship to collect 6568 throw events from 337 players of 24 national teams during 65 matches. ANOVA (F = 80.8, p < 0.01, η²  = 0.058) revealed that back players throws were significantly faster (24.3-26.5 m/s) than the rest of the players (22.2-22.5 m/s). Similarly, throws made from the central zone showed the highest speed values of 28.0 m/s (F = 43.2, p < 0.01, η²  = 0.050). In regard to the goal hit zone, higher speeds were shown for throws in side zones (24.9-26.8 m/s) compared to central zones (23.0-23.8 m/s) (F = 49.0, p < 0.01, η²  = 0.063). Effectiveness was higher in wing players (62%-64%) than the rest (45%-56%) (F = 30.9, p < 0.01, η²  = 0.023). Interestingly, the two higher velocity categories (>22 m/s) showed similar effectiveness (~60%), which were higher than the preceding slower category of 17-22 m/s (39.7%) (F = 175.1, p < 0.01, η²  = 0.074). Regarding court zones, throws from first-line zones (48%-60%), were more effective than from second line (38%-43%) (F = 13.1, p < 0.01, η²  = 0.016). With respect to the goal hit zone, effectiveness was higher in side (72%-77%) than in central zones (58%-64%) (F = 523.2, p < 0.01, η²  = 0.418). Coaches should improve the throwing velocity of players but also train at submaximal velocities (effective velocity) to save energy and reduce the potential risk of injury. Finally, training should be specialized according to the performance shown by players from different throwing zones, rather than from playing positions.

Defending in 4-4-2 or 5-3-2 formation? small differences in footballers' collective tactical behaviours

This study explored footballers' tactical behaviours, based on their position data, as an effect of two defending formations, 4-4-2 and 5-3-2, using an experimental approach. Sixty-nine youth footballers participated in this 11-versus-11 study, performing 72 trials of attack-versus-defence. Players' position data were tracked using a local positioning system, and processed to calculate measures of collective movement. This was supplemented by the analysis of passing networks. The results showed small differences between the two conditions. Compared to a 4-4-2 formation, defending in 5-3-2 reduced dispersion (-0.69 m,p=0.012), midfield-forward distance (-0.81 m, p=0.047), and defence-forward distance (-1.29 m, p=0.038); the consequent effects on attacking teams included reduced team widths (-1.78 m, p=0.034), reduced necessity for back-passes to the goalkeeper, and less connectivity in the passing network. The effects of the two defending formations seem to have the greatest impact on fullbacks of the attacking teams, since they were main contributors of the reduced team widths, received more passes, and had higher betweenness centrality in the right-back position during 5-3-2 defending. In summary, the present study potentially demonstrates how the underlying mechanisms in players' collective movements and passing behaviours show that the 5-3-2 is more conservatively defensive than the 4-4-2.

Artificial Intelligence Based Body Sensor Network Framework-Narrative Review: Proposing an End-to-End Framework using Wearable Sensors, Real-Time Location Systems and Artificial Intelligence/Machine Learning Algorithms for Data Collection, Data Mining and Knowledge Discovery in Sports and Healthcare

With the rising amount of data in the sports and health sectors, a plethora of applications using big data mining have become possible. Multiple frameworks have been proposed to mine, store, preprocess, and analyze physiological vitals data using artificial intelligence and machine learning algorithms. Comparatively, less research has been done to collect potentially high volume, high-quality 'big data' in an organized, time-synchronized, and holistic manner to solve similar problems in multiple fields. Although a large number of data collection devices exist in the form of sensors. They are either highly specialized, univariate and fragmented in nature or exist in a lab setting. The current study aims to propose artificial intelligence-based body sensor network framework (AIBSNF), a framework for strategic use of body sensor networks (BSN), which combines with real-time location system (RTLS) and wearable biosensors to collect multivariate, low noise, and high-fidelity data. This facilitates gathering of time-synchronized location and physiological vitals data, which allows artificial intelligence and machine learning (AI/ML)-based time series analysis. The study gives a brief overview of wearable sensor technology, RTLS, and provides use cases of AI/ML algorithms in the field of sensor fusion. The study also elaborates sample scenarios using a specific sensor network consisting of pressure sensors (insoles), accelerometers, gyroscopes, ECG, EMG, and RTLS position detectors for particular applications in the field of health care and sports. The AIBSNF may provide a solid blueprint for conducting research and development, forming a smooth end-to-end pipeline from data collection using BSN, RTLS and final stage analytics based on AI/ML algorithms.

Simulating Defensive Trajectories in American Football for Predicting League Average Defensive Movements

American football is an appealing field of research for the use of information technology. While much effort is made to analyze the offensive team in recent years, reasoning about defensive behavior is an emergent topic. As defensive performance and positioning largely contribute to the overall success of the whole team, this study introduces a method to simulate defensive trajectories. The simulation is evaluated by comparing the movements in individual plays to a simulated league average behavior. A data-driven ghosting approach is proposed. Deep neural networks are trained with a multi-agent imitation learning approach, using the tracking data of players of a whole National Football League (NFL) regular season. To evaluate the quality of the predicted movements, a formation-based pass completion probability model is introduced. With the implementation of a learnable order invariant model, based on insights of molecular dynamical machine learning, the accuracy of the model is increased to 81%. The trained pass completion probability model is used to evaluate the ghosted trajectories and serves as a metric to compare the true trajectory to the ghosted ones. Additionally, the study evaluates the ghosting approach with respect to different optimization methods and dataset augmentation. It is shown that a multi-agent imitation learning approach trained with a dataset aggregation method outperforms baseline approaches on the dataset. This network and evaluation scheme presents a new method for teams, sports analysts, and sports scientists to evaluate defensive plays in American football and lays the foundation for more sophisticated data-driven simulation methods.

Differences in situational power performance between playing positions in top level handball

ABSTRACT This study aimed to determine the differences in situational power performance between playing positions in handball. The following variables were analyzed: body height and weight, fastest shot, fastest sprint, highest jump, and average game time. The study sample comprised 412 handball players who participated at European championships. Backcourt players had the fastest shots and wing players presented the slowest shots among all playing positions. Wing players presented the fastest sprinting (29.09 km/h). Jumping performance showed the most diversity between the playing positions, and backcourt players jumped significantly higher than other players (16.76 cm).