Deep Learning-Based Object Detection, Localisation and Tracking for Smart Wheelchair Healthcare Mobility.

Louis Lecrosnier,Romain Rossi,Benoit Decoux,Redouane Khemmar,Naceur Kefi,Nicolas Ragot,Jean-Yves Ertaud

doi:10.3390/ijerph18010091

Abstract

This paper deals with the development of an Advanced Driver Assistance System (ADAS) for a smart electric wheelchair in order to improve the autonomy of disabled people. Our use case, built from a formal clinical study, is based on the detection, depth estimation, localization and tracking of objects in wheelchair’s indoor environment, namely: door and door handles. The aim of this work is to provide a perception layer to the wheelchair, enabling this way the detection of these keypoints in its immediate surrounding, and constructing of a short lifespan semantic map. Firstly, we present an adaptation of the YOLOv3 object detection algorithm to our use case. Then, we present our depth estimation approach using an Intel RealSense camera. Finally, as a third and last step of our approach, we present our 3D object tracking approach based on the SORT algorithm. In order to validate all the developments, we have carried out different experiments in a controlled indoor environment. Detection, distance estimation and object tracking are experimented using our own dataset, which includes doors and door handles.

Highlights

Object detection, recognition, localization and tracking are very important tasks in mobile robotics and computer vision applications
Before presenting our experiments for object detection and distance measurement, we need to define the different metrics that we use in the evaluation process
We have presented an object detection, depth estimation, location and tracking system for wheelchair healthcare smart mobility, for indoor environment, based on deep learning for object detection (YOLOv3), and various techniques and algorithms

Summary

Introduction

Recognition, localization and tracking are very important tasks in mobile robotics and computer vision applications. These processes are achieved through the use of different measurement sensors (camera, LIDAR, RADAR, etc.) and algorithms (filtering, object detection, pattern recognition, feature extraction, segmentation, classification, etc.). Through the ADAPT (“Assistive Devices for empowering disAbled People through robotic Technologies”, http://adapt-project.com) project, the detection and classification of objects which are specific to the wheelchair indoor environment, are a key issue for the safety navigation of the wheelchair. These objects are doors and door handles. We developed a specific dataset, and validated the complete approach through the ESIGELEC Autonomous Navigation

Objectives

Methods

Results

Conclusion