AI Revolutionizes Autonomous Wheelchairs

AI at the Service of Smart Wheelchairs

Users of wheelchairs with severe disabilities often navigate tight spaces with a dexterity that robotic systems struggle to match. A new wave of research on smart wheelchairs aims to bridge this gap through artificial intelligence. Promising results were presented in Anaheim, California, this month, where researchers are testing whether AI-powered systems can enhance user mobility.

Christian Mandel, a senior researcher at the German Research Center for Artificial Intelligence (DFKI) in Bremen, Germany, co-led a team with Serge Autexier to develop prototypes of electric wheelchairs. These wheelchairs are equipped with sensors designed to navigate rooms filled with potential obstacles. A new safety system has also been tested, integrating data from the wheelchair's sensors and sensors in the room, including color and depth cameras based on drones.

Mandel specifies that the team's smart wheelchairs operate in both semi-autonomous and autonomous modes. In semi-autonomous mode, the user drives using a joystick, while the autonomous mode is controlled by natural language commands, such as "Please take me to the coffee machine."

Advanced Technology for Precise Navigation

The experiments conducted by the researchers are part of a larger project called Reliable and Explainable Swarm Intelligence for People With Reduced Mobility (REXASI-PRO). Two identical smart wheelchairs were used, each equipped with two lidars, a 3D camera, odometers, user interfaces, and an embedded computer. The autonomous mode relies on the ROS2 Nav2 navigation system, using natural language input. The wheelchairs also utilize simultaneous localization and mapping (SLAM) and motion controllers to avoid obstacles.

One tested scenario involved the user pressing a button on the wheelchair's human-machine interface, speaking a command, and then confirming or rejecting the instruction through the same interface. Once the command was confirmed, the mobility device guided the user along a path to the destination, while the sensors detected obstacles and adjusted the device accordingly to avoid them.

Challenges of Accessibility and Reliability

Pooja Viswanathan, CEO and founder of Braze Mobility in Toronto, emphasizes that research in mobile assistive technology must also focus on the availability of these devices for everyday consumers. Cost remains a major barrier, as funding systems often do not support advanced intelligence without clear evidence of value and safety. Reliability is also a challenge, as a smart wheelchair must function in both ideal and chaotic everyday conditions. Users have different cognitive, motor, sensory, and environmental needs, so a one-size-fits-all solution does not apply.

Braze manufactures blind spot sensors for electric wheelchairs that detect obstacles in areas difficult for users to see. These sensors can be added to any wheelchair to transform it into a smart wheelchair, providing multimodal alerts to the user. This approach aims to support users rather than replace them.

Louise Devinge, a biomedical research engineer at IRISA (Institute for Research in Computer Science and Random Systems) in Rennes, France, notes that the increased complexity of smart wheelchairs requires more sensors, which necessitates careful management of communication and synchronization within the wheelchair system. More sensors and autonomy make it challenging to ensure robust performance in all real-world environments encountered by wheelchair users.

Future Perspectives for Smart Wheelchairs

Christian Mandel anticipates that smart wheelchairs could be ready for the consumer market within the next ten years. Pooja Viswanathan indicates that the REXASI-PRO system, although currently beyond the reach of existing technologies, is crucial in the long term. It represents the ambitious end of the smart wheelchair spectrum, with strengths in intelligent navigation, advanced sensing, and the effort to build a wheelchair capable of interpreting and responding to complex environments more autonomously.

From a research perspective, this type of work advances the field and takes seriously the importance of reliable and explainable AI. This is essential in any mobility technology where safety, reliability, and user trust are paramount.

Mandel, inspired by his beginnings in the field, recalls that he contributed to the development of a joystick-controlled smart wheelchair. However, he realized that even individuals with severe disabilities navigate very well through narrow passages, emphasizing that technology should complement users' abilities without underestimating them.

Researchers from DFKI presented their work at the CSUN Assistive Technology Conference in Anaheim, California, earlier this month.