The use of wearable devices in the assessment of basketball players’ athletic performance

Basketball is a high-intensity sport, which requires athletes to have good physical quality and technical level. The assessment of basketball players’ performance has been an important direction of basketball research [1-2]. By evaluating the athletic performance of basketball players, the problems and deficiencies of the athletes can be discovered in time, and targeted training can be carried out to improve the overall team’s level of play and performance, and the application of wearable devices can realize this goal [3-6].

With the rapid development of science and technology, smart wearable devices have quietly entered our lives. In the field of sports, these small devices are like a magical key that opens the door to excellence for athletes [7-10]. First of all, smart wearable devices can provide athletes with timely health feedback through real-time monitoring of heart rate, blood pressure, oxygen saturation and other physiological indicators, by identifying potential health problems and reminding athletes to take measures in a timely manner. This immediate feedback mechanism enables athletes to better understand their own physical condition, so as to arrange the training program more scientifically [11-14]. Secondly, smart wearable devices are able to record athletes’ trajectory, speed, strength and other data, and analyze them through algorithms. These data show every detail of the athletes in training. Through the in-depth mining of these data, athletes can discover their strengths and weaknesses, so as to develop a more accurate training program [15-18].

Currently there are relatively few studies related to wearable devices in basketball players’ performance evaluation, but there are relatively more studies in other aspects of basketball, such as training and physical fitness evaluation, but they show certain limitations. Literature [19] proposed a method for evaluating the training performance of basketball players based on wearable devices and machine learning algorithms, and revealed the practicality of the method by constructing an evaluation model. Literature [20] examines the application of wearable devices in the evaluation system of basketball technology and physical fitness, where athletes can better understand their own deficiencies by detecting the output data of smart devices, so as to target training to improve physical fitness and basketball technology. In literature [21], wearable devices are used to develop devices that can monitor the performance of basketball players and analyze the shooting posture, as well as a system that measures and analyzes the free throw movements of professional basketball players in real time. However, in literature [22], although wearable devices are an effective technology for analyzing sports performance, it suffers from problems such as high price, privacy and security, so a low-cost, privacy-protecting method for detecting and predicting the results of basketball shots was proposed and its feasibility was verified. Literature [23] combines deep learning algorithms and wearable devices to establish a feature extraction model of basketball data, constructs a basketball pose recognition method, realizes the technical analysis of basketball poses, and has an impetus to the field of recognition applications. Literature [24] aims to point out the importance of athlete recognition in the field of sports based on wearable devices through real-time monitoring of athletes’ physical status and the division of basketball movements, and emphasizes that this field has not yet received extensive attention in China at present.

From the above, it can be seen that the application of wearable devices in basketball has certain limitations and does not fully reflect the characteristics of wearable devices and their impact on sports and athletes. Literature [25] describes the impact of machine learning on player and team performance in different sports, as well as the positive impact of wearable technology on athletes, which is conducive to helping athletes understand their performance level for further improvement. Literature [26] provides an indication of the reliability of performance metrics based on wearable devices based on a literature review, as well as key considerations and future trends. Literature [27] examined the application of wearable devices in assessing athlete performance in sports activities with key metrics values sent to other electronic devices through the use of a wireless network and experimental data in order to illustrate the validity of this approach. Literature [28] highlights the potential of wearables and AI in sports and in the areas of injury prevention and training optimization for athletes. It reveals that the combined application of technologies facilitates the improvement of athletes’ performance and provides personalized care for their physical well-being. Literature [29] proposed a design method for a low-power stretchable wearable epidermal electronic device, which is described to have a preventive effect on serious health complications, injuries, and sudden death of athletes during training. Literature [30] developed a discussion on wearable devices based on a literature review, discussing their limitations in terms of accuracy, interpretability and cost. It also reveals that wearable devices can play an important role in improving the performance of athletes.

The article studies the sports performance evaluation system of smart wearable devices, and selects the electromyographic and myoacoustic signals within it for in-depth analysis. By extracting the athlete’s human EMG signal and myoacoustic signal, the real-time data monitoring of the athlete is carried out, so as to provide the smart wearable device to regulate the athlete’s sports status and intensity, and protect the athlete’s skeletal muscle injury. The performance of this paper’s method and the traditional electrode monitoring method for EMG signal monitoring is tested through shooting experiments. FMS functional indicators are selected to test the status of athletes’ skeletal muscle injuries and analyze the athletes’ sports performance before and after the use of smart wearable devices, so as to assess the effectiveness of smart wearable devices.

2

Methods for assessing athletic performance in smart wearable devices

2.1

Motion assessment methods for smart wearable devices

2.1.1

Smart Wearable Devices

Smart wearable devices are electronic devices with intelligent functions that can be worn or worn on a person’s body [31]. Smart wearable devices are mainly electronic devices that realize data interaction and data analysis through scientific hardware such as cameras, sensors, chips, etc., rich software installed on the end of the device and cloud control. The characteristics of smart wearable devices are mainly embodied in the aspects of intelligence, visualization and various models [32]. Among them, intelligence is mainly embodied in the equipped with intelligent functions, including voice assistant, big data analysis, etc., which can realize the intelligent analysis and feedback of the user’s instructions, and provide users with intelligent high-quality services.

Visualization features are mainly reflected in the intelligent wearable device can comprehensively record the user’s exercise data, and use the combination of charts and text to present the exercise situation, providing a basis for the user to exercise in a scientific and safe way. Personalized features reflect that smart wearable devices can provide personalized services according to the individual differences of different wearers, which facilitates the formation of targeted training programs and improves the scientific and targeted nature of sports training.

2.1.2

Assessment methodology

Most of the existing smart wearable strain sensors are based on the contact-resistance mechanism, i.e., the contact relationship change of the conductive microstructure is realized through various flexible sensing materials and corresponding microstructure design, so as to form the stretchability and strain-resistance change of the sensor. Commonly used sensing materials include carbon-based materials (e.g., carbon nanotubes, graphene, carbonized silk, carbon black, etc.) or metal nanowires (gold, silver, etc.), and the corresponding microstructures are spring-like structures, island-gap structures, and fish-scale-like structures.

According to the magnitude of strain, there are two main categories of flexible strain sensors to evaluate the range, including small muscle strain and large joint strain. Specifically: 1)

Assessment of small strains generated by facial muscles can be used to monitor mood changes.

2)

Evaluating small strains generated by arm, leg, and abdominal muscles can be used to monitor muscle fatigue, heart rate changes, and respiratory rhythms during exercise.

3)

Evaluating the strain at the joints and muscles of each body part of the athlete can be used for movement posture monitoring.

2.2

Evaluation of EMG Signals in Athletes

During the exercise process of basketball players, EMG signals are closely related to their exercise performance [33]. Through the surface electrode picks up the muscle surface electromyography signal for surface electromyography signal, surface electromyography signal of non-invasive and can well reflect the functional state of the muscle, in medicine, rehabilitation and other sports have been very good application. The human body’s exercise process is accompanied by the consumption of energy, with the consumption of energy, the human body’s muscles will appear a certain amount of fatigue, if muscle fatigue continues to exercise, it is very likely to be injured, in order to avoid injuries to the athletes, this paper designs how to use the surface EMG signal to know the fatigue of the human body’s muscles, so as to minimize the possibility of the injury caused by muscle fatigue for the athletes.

In recent years, time-frequency analysis methods have begun to attract people’s attention, and the commonly used time-frequency analysis methods for EMG signals are Short-Time Fourier Transform (STFT), Wavelet Transform (WT), Wigner Distribution, Empirical Modal Decomposition (EMD), etc., which are specifically introduced in the following.

2.2.1

Short-time Fourier transforms

The short-time Fourier transform, is based on the Fourier transform plus the window function, that is, based on the Fourier transform multiplied by the time function g(t), let the input signal is x(t), then the STFT calculation formula is as follows: (1) $S_{x}^{g} (τ, f) = \int_{- \infty}^{\infty} x (t) g^{*} (t - τ) e^{- i 2 π f t} d t$

2.2.2

Wigner distribution

Wigner distribution represents the distribution of signal energy in the time-frequency domain, and its definition domain has the advantages of inversion, unity, etc. It is more suitable for dealing with non-smooth signals, and let the input signal be x(t), then the formula of Wigner distribution is as follows: (2) $W_{P F_{i}} (t, f) = \int_{- \infty}^{\infty} x (t + τ / 2) x^{*} (t - τ / 2) \exp (- 2 π f t) d τ$

Where, τ represents the delay time, t is the time and f represents the signal frequency.

Since the transform of the Wigner distribution is bilinear, when the input signal has more components, the different components are prone to crossover, which in turn causes artifacts.

2.2.3

Wavelet transform

Wavelet decomposition is quite flexible in signal decomposition, Fourier transform has good localization ability in frequency domain but it has no localization ability in time domain whereas wavelet decomposition has time localization analysis ability. The properties of wavelet decomposition make it well used in signal processing. Let the acquired EMG signal be x(t), the wavelet decomposition is defined as follows: (3) $W (s, τ) = \int x (t) φ_{s, z}^{*} (t) d t$

where ϕ represents the wavelet basis function.

A wavelet sequence is obtained by stretching and translating the wavelet basis function as follows: (4) $φ_{s, τ} (t) = \frac{1}{\sqrt{s}} φ (\frac{t - τ}{s}) d t$

The functions of wavelet decomposition are obtained from equations (3) and (4) as follows: (5) $W (s, τ) = \frac{1}{\sqrt{s}} \int x (t) φ (\frac{t - τ}{s}) d t$

where s denotes the scaling factor.

2.2.4

Empirical Modal Decomposition

Empirical Modal Decomposition (EMD) overcomes the deficiencies of the Wigner distribution and wavelet transform by decomposing the signal into a number of intrinsic modal components (IMF) after EMD decomposition. Each IMF component represents a different frequency band range, so the original EMG signal can be decomposed into different frequency bands after EMD decomposition. The signal undergoes EMD decomposition to obtain multiple intrinsic modal components, and the IMF components must satisfy certain conditions: the function of subtracting the envelope average must satisfy that the number of extreme points and the number of zero crossing points are equal or differ by one. Calculate the average of the upper and lower envelopes to ensure that their mean value is zero. If it is not satisfied, the decomposition process is re-executed until it is satisfied. The EMD decomposition is specified as follows: 1)

Let the original signal be x(t) and find the extreme points of the signal, i.e., the local minima and local maxima.

2)

The upper envelope is formed by connecting the signals at the local extreme values determined in step 1) to form the envelope line, and the lower envelope is formed by connecting the signals at the local extreme values determined in step 1) to form the line.

3)

Let the average value of the upper and lower envelopes be m₁, and subtract this average value from the original signal to obtain a new data function labeled h₁, that is: (6) $h_{1} = x (t) - m_{1}$

According to the EMD decomposition principle at this point the signal being decomposed is a low frequency signal. Therefore, h₁ represents the high frequency component. At this time, it is necessary to determine whether h₁ belongs to the IMF component, if satisfied, then h₁ is the first IMF component after EMD decomposition, which is recorded as h₁ = c₁. 4)

If h₁ is not an IMF component, at this time, take h_i as the decomposition object and repeat steps 1)-3) until the condition of IMF is satisfied, marking the number of loops as k to obtain h_1k = h_1(k−1) − m_1k. Set the first MF component after decomposition as c₁, at this time c₁ = h_1k.

5)

Remove the first IMF component from the original signal, and since the EMD decomposition goes from high frequency to low frequency, thus a low frequency function is obtained, which is noted as r₁: (7) $r_{1} = x (t) - c_{1}$

At this point, the data function r₁ is taken as the decomposition object, and steps (1)-(5) are repeated to obtain the second IMF component of the EMD decomposition, which is noted as c₂. Performing the above steps n times, the n IMF components of the original EMG signal after EMD decomposition are obtained. The n IMF components of signal X(t) satisfy the following relationship: (8) ${\begin{matrix} r_{1} - c_{2} = r_{2} \\ ... \\ ... \\ r_{n - 1} - c_{n} = r_{n} \end{matrix}}$ 6)

The termination condition for the end of the loop is that r_n becomes a monotonic function. The decomposition process determines whether r_n is a monotonic function, and if so, terminates the loop, which gives: (9) $X (t) = \sum_{i = 1}^{n} c_{i} + r_{n}$

where c_i is each IMF component of the signal and r_n is the residual component.

The empirical modal decomposition method in time-frequency analysis is particularly suitable for nonsmooth, nonlinear signals, and its decomposition is adaptive, which can be well used for the evaluation of muscle functional state in combination with some other methods.

2.3

Assessment of muscle tone signals in athletes

2.3.1

Principle of muscle tone signal detection

The myoacoustic signal (MMG) reflects the activity state of human muscle contraction and is a low-frequency transverse vibration signal that records and quantifies skeletal muscle myofibers [34]. When a muscle undergoes autonomous or evoked movement conditions, the morphology of the muscle fibers changes, which generates a pressure wave within the muscle. MMG is a composite signal, which is formed by the joint action of motor units. Therefore, MMG contains information about the number of motor units and reflects the effect of the overall change in the size of the muscle fibers when they are involved in movement.MMG is a vibration signal whose frequency is mainly distributed in the low-frequency band, usually 0-100 Hz, and the energy is mainly concentrated in 2-35 Hz.

2.3.2

Musical Sound Signal Preprocessing and Feature Extraction

1)

Signal Filtering

MMG is a low-frequency weak biological signal, which is easily interfered by environmental noise, motion artifacts, or scattering noise during the acquisition process. Therefore, MMG needs to be filtered before analyzing it. The filtering processing methods for MMG in previous articles include Fourier transform-based denoising, wavelet denoising, wavelet packet filtering and empirical modal decomposition reconstruction denoising, etc. It can be seen that compared to traditional Fourier-variable signal processing methods, wavelet noise reduction and empirical modal decomposition reconstruction are more effective in processing, but the computation is also more complicated.

The Butterworth filter is the largest flat filter, which has the advantages of good stability, simple computation, and a small burden on system performance. The Butterworth filter can better remove interference noise from the MMG and retain the main signal, providing a better signal for subsequent MMG processing.

2)

Feature Extraction

After the MMG is filtered and processed, feature extraction is needed to select the feature parameters related to muscle strength as the input parameters of the muscle strength estimation model. Feature extraction usually uses time-domain features and frequency-domain features. The statistical properties of MMG in the time domain can be reflected by time-domain features, which can be calculated directly from the MMG sequence. The method is simple to compute, and the commonly used time-domain features of the myotonic signal are integral myotonic value (IMMG), absolute mean amplitude (MAV), mean square error (VAR), and root mean square (RMS). The expressions are as follows: (1)

IMMG, which is the integration of the time series of MMG, is expressed as: (10) $IMMG = \frac{\sum_{i = 1}^{N} | x_{i} |}{N}$

(2)

MAV, is the absolute mean value of the MMG amplitude, expressed as: (11) $MAV = \frac{1}{N} \sum_{i = 1}^{N} | x_{i} |$

(3)

VAR, which indicates the extent to which the MMG deviates from the mean, is expressed as: (12) $VAR = \frac{1}{N - 1} \sum_{i = 1}^{N} {(x_{i} - \bar{x})}^{2}$

(4)

RMS, characterizes the amplitude value at the maximum probability density of the signal and is expressed as: (13) $RMS = \sqrt{\frac{1}{N} \sum_{i = 1}^{N} x_{i}^{2}}$

where Equation x represents the ird data point of the myotonic signal x, N is the total number of data, and $\bar{x}$ is the average value.

The frequency domain features characterize the trend of the signal on the power spectrum, and the frequency domain analysis of MMG requires obtaining the power spectrum by Fourier transform, calculating the power spectral density, and obtaining the frequency domain features of the signal by statistical methods. The commonly used frequency domain features of MMG are mean power frequency (MPF), median frequency (MDF), mean power (MP), and root mean square of power spectrum (rmsPS). The expressions are as follows: (1)

MPF, reflecting the average frequency of the MMG, is expressed as: (14) $MPF = \sum_{j = 1}^{M} f_{j} P_{j} / \sum_{j = 1}^{M} P_{j}$ $${\rm{MPF}} = \mathop \sum \limits_{j = 1}^M {f_j}\>{P_j}/\mathop \sum \limits_{j = 1}^M {P_j}$$

(2)

MDF, is the frequency value corresponding to half of the energy spectrum and is expressed as: (15) $\sum_{j = 1}^{MDF} P_{j} = \sum_{j = MDF}^{M} P_{j} = \frac{1}{2} \sum_{j = 1}^{M} P_{j}$

(3)

MP, which reflects the average energy of the signal, is expressed as: (16) $MP = \frac{1}{M} \sum_{j = 1}^{M} P_{j}$

(4)

rmsPS, is the root mean square calculated for the power spectrum with the expression: (17) $rmsPS = r m s (P)$

where f_j is the jnd frequency value of the power spectrum, P_j is f_j the corresponding power spectral density value, and M is the total number of frequency values.

3

Applications in the assessment of basketball players’ sports performance

3.1

Evaluate performance analysis

Taking electromyography signals as an example, we test the results of the electromyography signal evaluation methods in this paper on each evaluation index, and compare them with the traditional electrode evaluation methods to judge the performance of each muscle signal detection method.

3.1.1

Shooting tests

At the same time period (14:00-18:00) on five different days, the experimental subjects were made to perform the shooting test, which included two kinds of movement performances, namely, fixed-point shooting and mobile shooting. The test results are shown in Table 1, where IEMG and RMS represent the integral EMG value and root mean square EMG value, respectively. It can be seen that the IEMG as well as the RMS indexes of EMG signals of the athletes during the fixed-point shooting were greater than the corresponding values of the mobile shooting, and the biceps force in the fixed-point shooting test was greater. When performing mobile shooting, the shooting movement of lateral movement is not well controlled, and there will be situations such as the basketball being covered by the body, the wire moving too much, and the palm tilting during the sustained force, which leads to a large fluctuation of the signal intensity and affects the subsequent evaluation of the EMG signals. Comparatively speaking, the biceps force is small and the fluctuation of signal intensity is small during fixed-point shooting, which helps to reduce the influence of movement on the EMG signals. The signal-to-noise ratios (SNR) of the five groups of shooting tests were calculated to be 18.67±0.89, 18.78±0.79, 19.15±0.43, 18.69±0.62, and 18.66±0.58, respectively, and the p-value results of the t-test of these five groups of data in two pairs were all greater than 0.05, which indicated that there was no significant difference in the SNRs of the five groups of tests under this maneuver. Therefore, the EMG signals measured by the shooting test were stable over time.

Table 1.

Analysis of shooting test results

Time period	Spot shot basketball		Moving shot basketball
Time period	IEMG (V)	RMS (V)	IEMG (V)	RMS (V)	SNR
1	0.433±0.039	0.103±0.015	0.251±0.013	0.064±0.001	18.67±0.89
2	0.405±0.048	0.106±0.016	0.247±0.025	0.066±0.003	18.78±0.79
3	0.386±0.018	0.092±0.006	0.242±0.016	0.064±0.001	19.15±0.43
4	0.347±0.021	0.091±0.005	0.233±0.021	0.060±0.003	18.69±0.62
5	0.342±0.025	0.088±0.007	0.236±0.017	0.061±0.001	18.66±0.58

Frequency domain analysis was performed on the shooting test, and the results are shown in Table 2, where MPF and MF represent the mean power frequency and median frequency of EMG signals, respectively. As can be seen from the table, MPF and MF values in the existence of certain fluctuations, but the five groups of frequency characteristics of the two two t-test, the average power frequency (MPF) and the median frequency (MF) of the p-value results are greater than 0.05, indicating that the five groups of tests under this action frequency domain characteristics are not significantly different. Therefore, the performance of the EMG monitoring application was evaluated later in this paper using the fixed-point shooting maneuver.

Table 2.

Analysis of frequency domain characteristics of shot

Time period	MPF (Hz)	MF (Hz)
1	58.45±1.16	50.63±2.42
2	57.89±2.05	49.15±2.03
3	56.15±2.08	48.33±2.51
4	57.05±1.22	49.44±2.24
5	56.84±1.24	48.51±2.08

3.1.2

Comparison of EMG monitoring performance

Spot shooting tests were performed on 10 occasions on the same day for this paper’s EMG signal detection method and the traditional electrode monitoring method, and the monitoring performance was evaluated by the RMS, signal-to-noise ratio (SNR), and the mean power frequency (MPF) of the signals, and the results are shown in Table 3. By comparing the RMS of the signals, it can be seen that the EMG signal strength monitored by this paper’s method (0.107) is higher than that of the conventional electrodes (0.056). This is mainly because the smart wearable device applied in this paper’s monitoring method has a larger conductive area than traditional electrodes, and more muscle fibers are monitored, thus the acquired EMG signal is strong. By comparing the signal-to-noise ratio of the EMG signals, it can be seen that the EMG monitoring performance of this paper’s method is better than that of the traditional electrode, and the t-test of the two shows that the p-value result is 0.008, which is smaller than 0.05, indicating that there is an obvious difference in their EMG monitoring performance. The average power frequency of EMG signals monitored by the method of this paper (67.45) and the traditional electrodes (57.56) were subjected to t-test in the table. The p-value results are 0.006, which is less than 0.05, indicating that there is a significant difference in the average power frequency of EMG signals between this paper’s monitoring method and traditional electrodes. Therefore, the monitoring method in this paper is significantly better than the performance of conventional electrodes for EMG signal monitoring.

Table 3.

Results of electromyography monitoring performance test

	Traditional electrode monitoring method	Ours	p
RMS (V)	0.056±0.008	0.107±0.042	0.012
SNR	18.62±0.71	25.14±2.54	0.008
MPF (Hz)	57.56±1.26	67.45±2.85	0.006

3.2

Analysis of monitoring effectiveness

This section of the experiment aims to demonstrate the effect of smart wearable devices on the assessment of sports performance of basketball players. 9 amateur basketball players (No. A~I) from Q City Amateur Sports School were selected as the research subjects, and they were randomly divided into two groups (experimental group and control group) to conduct sports experiments for 2 months. The FMS functional screening was chosen as an index to reflect the athletic performance status of basketball players.

3.2.1

Smart Wearables Pre-training FMS Test Results

The results of the FMS test before wearing smart devices for training are shown in Table 4. As can be seen in Table 4, the highest score of the nine adolescent athletes was 16 and the lowest was 12, with an overall mean score of 14.78. A total of six athletes showed painful conditions in the trunk stability and shoulder mobility tests and generally scored poorly in this test item of trunk stability. This indicates that they do not have enough control over their bodies in an unstable state.

Table 4.

FMS test results at the beginning of training experiment with wearable smart equipment

	A	B	C	D	E	F	G	H	I
Squat test	2	3	3	2	3	2	3	3	3
Stepping test	3	2	2	2	2	1	2	3	3
Straight row squat	2	2	3	3	3	2	2	2	3
Shoulder flexibility	3	2	0	3	0	3	1	2	0
Straight leg lift	2	3	2	3	3	2	1	1	2
Torso stability	2	0	2	0	2	3	0	3	2
Rotation stability	2	2	3	2	2	3	3	2	1
Total	16	14	15	15	15	16	12	16	14
Mean	2.29	2.00	2.14	2.14	2.14	2.29	1.71	2.29	2.00

3.2.2

FMS test results in smart wearable device training

In the middle of the experiment (1 month), 9 athletes and retrograde FMS indicators were tested again and the results are shown in Table 5. From Table 5 we can see that the highest FMS score is 18, the lowest score is 16, and the average score is 17. It is 2.22 points higher than the score before wearing the smart device training, and the number of people with less than 14 points is zero. Two of the youth athletes who felt pain in the shoulder flexibility test and two of the three youth athletes who felt pain in the trunk stability test prior to the test were already pain free during this test, indicating that their old injuries or strains had improved during the month of training with the smart device. In the trunk stability test, the number of people who improved was higher. The average score of the trunk stability test before the test was 1.56, and the average score of the trunk stability test is now 2.22, which is an increase of 0.66 points, or 42%. This indicates that shoulder flexibility and trunk stability improved more in the junior long jumpers after 1 month of training with the smart device.

Table 5.

FMS test results in the middle of the training experiment with wearable smart equipment

	A	B	C	D	E	F	G	H	I
Squat test	3	3	3	2	3	2	3	3	3
Stepping test	3	3	3	2	2	2	2	3	3
Straight row squat	3	2	3	3	3	2	3	2	3
Shoulder flexibility	3	2	2	3	1	3	3	2	0
Straight leg lift	2	3	2	3	3	2	1	1	2
Torso stability	2	1	3	0	3	3	2	3	3
Rotation stability	2	3	2	3	2	3	3	3	2
Total	18	17	18	16	17	17	17	17	16
Mean	2.57	2.43	2.57	2.29	2.43	2.43	2.43	2.43	2.29

The FMS data shows that basketball players have improved more in controlling their bodies. Among the FMS test items, the three items of linear lunge squat, trunk stability and rotational stability were improved by a larger number of people, which corresponds to the analysis mentioned in the previous section that the injury-prone parts of basketball sports are mostly concentrated in the lower limbs. It reflects that smart wearable devices are targeted and effective for improving the sports performance of basketball players.

3.2.3

FMS test results after smart wearable device training

At the end of the experiment, the FMS indicators of 9 adolescent athletes were measured and the results are shown in Table 6. From Table 6, we can see that the highest score among the 9 adolescent athletes is 20 and the lowest score is 18.The average score of the 9 people is 19.33, which is 31% and 14% higher than the average score of the first and second tests.Overall, the smart wearable devices have made a lot of progress in the training of adolescent basketball players. It shows that most of the youth athletes have improved their body stability and muscle control by wearing smart devices.

Table 6.

FMS test results in the end of the training experiment with wearable smart equipment

	A	B	C	D	E	F	G	H	I
Squat test	3	3	3	2	3	2	3	3	3
Stepping test	3	3	3	2	3	3	2	3	3
Straight row squat	3	3	3	3	3	2	3	3	3
Shoulder flexibility	3	3	2	3	3	3	3	2	3
Straight leg lift	2	3	2	3	3	3	2	3	2
Torso stability	3	2	3	2	3	3	2	3	3
Rotation stability	3	3	3	3	2	3	3	3	3
Total	20	20	19	18	20	19	18	20	20
Mean	2.86	2.86	2.71	2.57	2.86	2.71	2.57	2.86	2.86

In this FMS data, it can be seen that improvements in the two indicators concerning trunk stability and rotational stability are evident in adolescent basketball players, and improvements in these areas mean that adolescent basketball players are significantly less likely to experience skeletal muscle injuries.

3.2.4

Comparison of FMS test results before, during, and after training

The results of the three FMS tests before, during and after the experiment were compared and the results are shown in Table 7. As can be seen in Table 7, the results of the three tests were significantly different (P < 0.05), in which the average score of the nine adolescent basketball players improved by 4.33 points after 2 months of training. The FMS score values of every adolescent athlete saw an improvement, with the most noticeable increase of 6 points and the least noticeable increase of 3 points. It is possible to see an improvement in the physical health status of adolescent basketball players who have been training with smart devices for the last 2 months. The data from these changes proves that smart wearable devices can improve the athletic performance of basketball players.

Table 7.

Comparison of FMS test results

Test object	Beginning	Middle	End
A	16	18	20
B	14	17	20
C	15	18	19
D	15	16	18
E	15	17	20
F	16	17	19
G	12	17	18
H	16	17	20
I	14	16	20
Mean	14.78	17	19.33
t	4.252
p	0.005

4

Conclusion

The article analyzes the evaluation of human electromyography signal and myoacoustic signal by the sports performance evaluation method of smart wearable device, so as to grasp the human skeletal muscle data during exercise, and regulate it through the wearable device to protect athletes from skeletal muscle injury. 1)

In the shooting test, the RMS value of the EMG signal monitored by this paper’s EMG signal monitoring method is 0.107, which is higher than that of the traditional electrode monitoring method, which is 0.056, and the EMG signal collected by this paper’s monitoring method is stronger than that of the traditional electrode method. The signal-to-noise ratio (SNR) of the EMG signal of this paper is 25.14, and the mean power frequency (MPF) is 67.45, which are better than that of the traditional electrode, and the p-value results of the three indexes are less than 0.05, so that this paper’s method has a better performance of monitoring the EMG signal.

2)

The mean score of FMS test of the experimental subjects before training with the smart wearable device was 14.78, and 6 subjects had pain in the trunk stability and shoulder flexibility tests. After 1 month of using the smart wearable device, the mean score improved to 17 and the number of subjects experiencing pain in the trunk stability and shoulder mobility tests decreased to 2 subjects. The mean FMS score after 2 months of training using the smart wearable device was 19.33, and pain was present in 0 of the trainers. The p-value of the results of the three tests was less than 0.05, which represents a significant difference. Smart wearable devices are effective in enhancing the athletic performance of basketball players.

Langue:: Anglais

Périodicité:: 1 fois par an
Sujets de la revue:: Sciences de la vie, Sciences de la vie, autres, Mathématiques, Mathématiques appliquées, Mathématiques générales, Physique, Physique, autres

RSS Feed de la revue

The use of wearable devices in the assessment of basketball players’ athletic performance

Zhaozhao Wu

Minxiao Xu

Chaoyi Qu

Publié en ligne: 24 mars 2025

Reçu: 30 oct. 2024

Accepté: 06 févr. 2025

DOI: https://doi.org/10.2478/amns-2025-0754

Mots clésSmart wearable devices, Electromyographic signals, Myoacoustic signals, Motor performance assessment

© 2025 Zhaozhao Wu et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Mots clés
Smart wearable devices, Electromyographic signals, Myoacoustic signals, Motor performance assessment