Expanding the field of view
Updated: Apr 4, 2022
This article is written by Roman Adámek who followed a three-month internship with the Department of Systems and Control under the supervision of Dr Marvin Bugeja and Prof Simon Fabri
For many people, systems and control engineering might be terms that are hard to imagine and connect with something tangible, even though we come in contact with its products and results daily. One of the areas which belong to the field of system and control engineering is robotics, and in the case of this project, the focus is on assistive robotics. Assistive robots are believed to be the future of modern health care and nursing. A good example of an assistive robot is the smart wheelchair designed and developed by the Department of System and Control Engineering of the University of Malta. And I’m very grateful that I had the opportunity to spend five months working on its enhancement and development as an intern.
Assistive robots are believed to be the future of modern health care and nursing.
The smart wheelchair project was initiated following talks between an wheelchair user and Prof. Ing. Kenneth Camilleri who then, together with Dr Ing. Marvin Bugeja and Prof. Simon Fabri worked to secure funds for the project to purchase a motorised wheelchair as well as a computer, sensors and electronics required to smarten-up the wheelchair. This process was carried out in the form of an undergraduate final year project carried out by Matthew Aquilina, worked under the supervision of Dr Ing. Marvin Bugeja and Prof. Ing. Simon Fabri.
The aim was to convert a standard powered wheelchair into a low-cost smart wheelchair having a set of assistive features to make its use easier, and therefore, more accessible to a wider audience. What makes the wheelchair smart is that it is equipped with sensors enabling it to detect obstacles in its environment and take appropriate actions to ascertain both safety and comfort for its user. For instance, the smart wheelchair can be controlled only with minimal input commands from the user (such as the desired direction) while it avoids obstacles and follows the designated direction autonomously. Another way of using the smart wheelchair is to select a destination on a map and let it transport you all by itself. The wheelchair will select a suitable path and avoid any obstacles on its way.
Figure 1: A planar lidar can perceive obstacles only in one plane and cannot avoid obstacles lower or above that plane.
My goal during the internship was, among other things, to improve the obstacle avoidance capability of the wheelchair. In its original state, it was equipped only with a planar laser-range-finder (LIDAR). Therefore, it was able to perceive obstacles only in one plane as shown in Figure 1, and could not avoid obstacles on the floor, tables, cats, etc. for example.
To tackle this problem, we decided to expand its field of view by equipping the wheelchair with an RGBD camera. In contrast to a regular RGB camera, this depth camera can see how far the obstacles are. Information from both sensors (LIDAR and RGBD camera) is then combined (fused) into a local costmap which is used as an input into an obstacle avoidance algorithm which was also enhanced and improved during this internship. The improved obstacle avoidance algorithm is based on the Dynamic Window Approach (DWA) planner combined with Vector Filed Histogram (VFH) algorithm. The wheelchair is also constantly performing Simultaneous Localization and Mapping (SLAM) to create and update the map of its environment, and can therefore adapt to any map changes (such as new obstacles) that it encounters. Several available SLAM algorithms were assessed in a series of experimental tests and one of them, an ROS algorithm named RTAB-Map, turned out to be the best choice for this application. It is a graph-based SLAM algorithm which uses incremental appearance-based loop closure detector. Further improvements were made in the GUI and start-up procedure of the smart wheelchair, with the main intention being that of making it more user-friendly.
My internship at the Department of System and Control Engineering expanded not only the field of view for the wheelchair but for me as well. Department provides excellent facilities for research and development. And this way, I would like to thank all the staff for being so welcoming and helpful whenever needed and letting me experience Malta from a completely different perspective.