Event-Driven On-Sensor Locomotion Mode Recognition Using a Shank-Mounted IMU with Embedded Machine Learning for Exoskeleton Control

Mohammadsaleh Razmi; Iman Shojaei

cs.ROFeb 24, 2026

Event-Driven On-Sensor Locomotion Mode Recognition Using a Shank-Mounted IMU with Embedded Machine Learning for Exoskeleton Control

Authors: Mohammadsaleh Razmi, Iman Shojaei

Organizations: Align Technology, Inc., San Jose, CA 95134, USA

Abstract

This work presents a wearable human activity recognition (HAR) system that performs real-time inference directly inside a shank-mounted inertial measurement unit (IMU) to support low-latency control of a lower-limb exoskeleton. Unlike conventional approaches that continuously stream raw inertial data to a microcontroller for classification, the proposed system executes activity recognition at the sensor level using the embedded Machine Learning Core (MLC) of the STMicroelectronics LSM6DSV16X IMU, allowing the host microcontroller to remain in a low-power state and read only the recognized activity label from IMU registers. While the system generalizes to multiple human activities, this paper focuses on three representative locomotion modes - stance, level walking, and stair ascent - using data collected from adult participants. A lightweight decision-tree model was configured and deployed for on-sensor execution using ST MEMS Studio, enabling continuous operation without custom machine learning code on the microcontroller. During operation, the IMU asserts an interrupt when motion or a new classification is detected; the microcontroller wakes, reads the MLC output registers, and forwards the inferred mode to the exoskeleton controller. This interrupt-driven, on-sensor inference architecture reduces computation and communication overhead while preserving battery energy and improving robustness in distinguishing level walking from stair ascent for torque-assist control.

Figures & tables

Figure 1: Shank-mounted SensorTile.box placement and coordinate definition used for data collection.

Figure 2: Data-collection activities using a shank-mounted IMU: ( a ) stance, ( b ) level walking on a treadmill, and ( c ) stair ascent on a step climber.

Figure 3: Tri-axial accelerometer signals (10 s) for stance, walking, and stair ascent. Columns: stance, walk, stairsUp. Rows: $a_{x}$ , $a_{y}$ , $a_{z}$ . All subplots in each row share the same y-axis limits for direct comparison.

Figure 4: Tri-axial gyroscope signals (10 s) for stance, walking, and stair ascent. Columns: stance, walk, stairsUp. Rows: $\omega_{x}$ , $\omega_{y}$ , $\omega_{z}$ . All subplots in each row share the same y-axis limits for direct comparison.

Figure 5: Experimental protocol snapshots for real-time evaluation: ( a ) stance, ( b ) level walking, and ( c ) stair ascent. The MLC configuration was pre-loaded on the IMU and the predicted class was read from the MLC output registers during the trial.

Mode	Start (s)	End (s)	Duration (s)
Stance	0.00	5.39	5.39
Walk	5.39	8.62	3.23
StairsUp	8.62	17.24	8.62

Table 1: Example on-sensor mode timeline from the logged MLC output ( dec_tree_out_1 ).

Figure 6: Example real-time experiment showing raw IMU signals and on-sensor inference output. ( a ) Tri-axial accelerometer, ( b ) tri-axial gyroscope, and ( c ) MLC decision-tree predicted mode over time (8=stance, 0=walk, 4=stairsUp).

Explore similar work

Mar 16, 2026

User-Tailored Learning to Forecast Walking Modes for Exosuits

Assistive robotic devices, like soft lower-limb exoskeletons or exosuits, are widely spreading with the promise of helping people in everyday life. To make such systems adaptive to the variety of users wearing them, it is desirable to endow exosuits with advanced perception systems. However, exosuits have little sensory equipment because they need to be light and easy to wear. This paper presents a perception module based on machine learning that aims at estimating 3 walking modes (i.e., ascending or descending stairs and walking on level ground) of users wearing an exosuit. We tackle this perception problem using only inertial data from two sensors. Our approach provides an estimate for both future and past timesteps that supports control and enables a self-labeling procedure for online model adaptation. Indeed, we show that our estimate can label data acquired online and refine the model for new users. A thorough analysis carried out on real-life datasets shows the effectiveness of our user-tailored perception module. Finally, we integrate our system with the exosuit in a closed-loop controller, validating its performance in an online single-subject experiment.

Gabriele Abbate, Enrica Tricomi, Nathalie Gierden +3

Dalle Molle Institute for Artificial Intelligence (IDSIA), USI-SUPSI, Lugano, Switzerland · Munich Institute for Robotics and Machine Intelligence (MIRMI), Department of Computer Engineering, School of Computation, Information and Technology, Technical University of Munich (TUM), Munich, Germany.

Assistive Robotics Wearable Sensing Online Learning

Ask PDF

Oct 13, 2026

Tread: A Visuo-Inertial Framework for Terrain and Locomotion Context-Aware Gait-Phase and Speed Estimation in Wearable Robotics

A wearable assistive robot must continuously estimate terrain, locomotion mode, gait phase, and walking speed. These quantities are commonly addressed by separate, terrain-specific models, with datasets rarely collected under a common protocol. This paper presents a visuo-inertial framework that connects these estimates through a causal pipeline using a chest-mounted camera and four lower-limb inertial measurement units (IMUs). Anticipatory Terrain Perception (ATP) classifies incoming terrain with 96.3% accuracy and anticipates terrain transitions by 1.46 s on average. Terrain-Aware Locomotion Mode Identification (TALMI) achieves 93.7% and 98.8% accuracy for its level-ground and stairs experts respectively, with mean transition latencies being below 200 ms. The relevant TALMI posterior is retained as a Probabilistic Locomotion Context (PLC) vector and provided to a shared Gait Kinematic Encoder (GKE), which supports 50-Hz bilateral gait-phase estimation and 2-Hz walking-speed estimation. Leave-one-subject-out evaluation yields gait-phase RMSEs of 1.75%, 2.97%, and 4.55% for level walking, stair ascent, and stair descent respectively while estimating speed over 3.5-5.5 km/h with a RMSE of 0.218 km/h. All inertial sensing based models use sensor-fused Euler angles as input rather than raw accelerometer and gyroscope signals, reducing dependence on sensor mounting variation.

Debadrata Sarkar, Akash Bachhar, Aman Arora

Human Centred Robotics and Cybernetics Group, CSIR-Central Mechanical Engineering Research Institute, Durgapur 713209, India

Terrain-Aware Robot Locomotion Gait Analysis Wearable Sensing

Ask PDF

Apr 28, 2026

Modular Retrieval-Augmented Generalization for Human Action Recognition

Inertial Measurement Unit (IMU)-based Human Activity Recognition (HAR) aims to interpret and classify user behaviors from temporal motion signals. Recently, deep learning frameworks have advanced this task by learning and extracting discriminative spatiotemporal representations, significantly improving recognition performance. However, IMU-based HAR still faces several critical challenges, particularly limited training samples and static knowledge utilization, both of which severely hinder its large-scale deployment. In this paper, we introduce MoRA, the first Retrieval-Augmented Module specifically designed for motion series. It can be flexibly integrated into any existing HAR model, enhancing recognition performance while maintaining inference efficiency. To address issues such as information redundancy in retrieval results and rigid fusion strategies, we propose an uncertainty-adaptive fusion unit within MoRA. This unit leverages previous physical knowledge from IMU signals to dynamically adjust the fusion strategy between original outputs and retrieved information, enabling more robust recognition. Extensive experiments on ten real-world datasets demonstrate that MoRA significantly improves the performance of existing IMU-based HAR models, consistently delivering stable and effective gains. The source code of MoRA is available at: https://github.com/liavonpenn/mora.

Peng Liao, Shangsong Liang, Lin Chen +1

School of Computer Science and Engineering, Sun Yat-sen University · Engineering Research Centre of Applied Technology on Machine Translation and Artificial Intelligence, Macao Polytechnic University

Human Activity Recognition

Ask PDF