Alpine Skiing Activity Recognition Using Smartphone’s IMUs

Self Cite

Learning underlying patterns from sensory data is crucial in the Human Activity Recognition (HAR) task to avoid poor generalization when coping with unseen data. A key solution to such an issue is representation learning, which becomes essential when input signals contain activities with similar patterns or when patterns generated by different subjects for the same activity vary. To address these issues, we seek a solution to increase generalization by learning the underlying factors of each sensor signal. We develop a novel multi-channel asymmetric auto-encoder to recreate input signals precisely and extract indicative unsupervised futures. Further, we investigate the role of various activation functions in signal reconstruction to ensure the model preserves the patterns of each activity in the output. Our main contribution is that we propose a multi-task learning model to enhance representation learning through shared layers between signal reconstruction and the HAR task to improve the robustness of the model in coping with users not included in the training phase. The proposed model learns shared features between different tasks that are indeed the underlying factors of each input signal. We validate our multi-task learning model using several publicly available HAR datasets, UCI-HAR, MHealth, PAMAP2, and USC-HAD, and an in-house alpine skiing dataset collected in the wild, where our model achieved 99%, 99%, 95%, 88%, and 92% accuracy. Our proposed method shows consistent performance and good generalization on all the datasets compared to the state of the art.

“…These public datasets have been widely used in the HAR literature [ 65 ]. The fifth dataset is an alpine skiing dataset [ 66 ].…”

Section: Approach and Methodsmentioning

confidence: 99%

“…The importance of this dataset is that some sessions are recorded uncontrolled which makes it appropriate for a real case scenario test. Please refer to [ 66 ] for detailed information. It is worth mentioning that some of the skiers performed skiing activities in different locations.…”

Section: Approach and Methodsmentioning

confidence: 99%

Robust Feature Representation Using Multi-Task Learning for Human Activity Recognition

Azadi,

Haslgrübler,

Anzengruber-Tanase

et al. 2024

Self Cite

“…IMU-based sports activity recognition is one of the main application fields for HAR studies, as summarized in Tables 2 and 3. It has already been proven for a variety of different sports, such as running [37][38][39], ball sports [36,[40][41][42][43][44], winter sports [45,46], sports for the disabled [47], or fitness [48][49][50][51], that activity recognition algorithms are capable of detecting specific activities tied to these sports based on IMU data as input. Other studies that incorporated IMUs focus rather on the athletes' performance [52][53][54][55] or gait estimation [56,57] than activity recognition.…”

Section: Related Work On Sports Studiesmentioning

confidence: 99%

Hang-Time HAR: A Benchmark Dataset for Basketball Activity Recognition Using Wrist-Worn Inertial Sensors

Hoelzemann

Romero

Bock

et al. 2023

We present a benchmark dataset for evaluating physical human activity recognition methods from wrist-worn sensors, for the specific setting of basketball training, drills, and games. Basketball activities lend themselves well for measurement by wrist-worn inertial sensors, and systems that are able to detect such sport-relevant activities could be used in applications of game analysis, guided training, and personal physical activity tracking. The dataset was recorded from two teams in separate countries (USA and Germany) with a total of 24 players who wore an inertial sensor on their wrist, during both a repetitive basketball training session and a game. Particular features of this dataset include an inherent variance through cultural differences in game rules and styles as the data was recorded in two countries, as well as different sport skill levels since the participants were heterogeneous in terms of prior basketball experience. We illustrate the dataset’s features in several time-series analyses and report on a baseline classification performance study with two state-of-the-art deep learning architectures.

“…Ref. [ 21 ] used the IMU found inside a smartphone to track a skier through uncontrolled, in-field testing. The in-field sensing systems mentioned above provided valuable insight into skier performance, but were not able to investigate the dynamic motion of the ski equipment.…”

Section: Introductionmentioning

confidence: 99%

Distributed IMU Sensors for In-Field Dynamic Measurements on an Alpine Ski

Beuken,

Priest,

Hainsworth

et al. 2024

Modern ski design is an inherently time-consuming process that involves an iterative feedback loop comprised of design, manufacturing and in-field qualitative evaluations. Additionally consumers can only rely on qualitative evaluation for selecting the ideal ski, and due to the variation in skier styles and ability levels, consumers can find it to be an inconsistent and expensive experience. We propose supplementing the design and evaluation process with data from in-field prototype testing, using a modular sensor array that can be ported to nearly any ski. This paper discusses a new distributed Inertial Measurement Unit (IMU) suite, including details regarding the design and operation, sensor validation experiments, and outdoor in-field testing results. Data are collected from a set of spatially distributed IMUs located on the upper surface of the ski. We demonstrate that this system and associated post-processing algorithms provide accurate data at a high rate (>700 Hz), enabling the measurement of both structural and rigid ski characteristics, and are robust to repetitive testing in outdoor winter conditions.