For many wheelchair users, navigating tight spaces isn’t a technological challenge—it’s a honed skill. But a growing wave of research, including findings presented earlier this month at the CSUN Assistive Technology Conference in Anaheim, California, is exploring whether artificial intelligence can further enhance mobility for those with severe disabilities. The goal isn’t to replace human skill, but to augment it, creating “smart wheelchairs” capable of navigating complex environments with greater ease and independence.
Researchers at the German Research Center for Artificial Intelligence (DFKI) in Bremen, Germany, are at the forefront of this effort. Led by Christian Mandel, a senior researcher, and his colleague Serge Autexier, the team has developed prototype sensor-equipped electric wheelchairs designed to tackle real-world obstacles. Their approach integrates data from multiple sources, including lidar, 3D cameras, and even drone-based color and depth sensors, to create a comprehensive understanding of the surrounding environment. This technology aims to address a critical need for more sophisticated assistive devices.
The DFKI team’s function centers around two modes of operation: semiautonomous and fully autonomous. In semiautonomous mode, the user retains control via a traditional joystick, while the AI provides assistance with obstacle avoidance and path planning. Fully autonomous operation, however, allows users to simply state their destination – “Please drive me to the coffee machine,” as Mandel explains – and let the wheelchair navigate the route independently. This relies on the open-source ROS2 Nav2 navigation system and simultaneous localization and mapping (SLAM) technology, allowing the wheelchair to build and maintain a map of its surroundings.
The research, part of a larger project called REXASI-PRO (Reliable and Explainable Swarm Intelligence for People With Reduced Mobility), involved extensive testing with two identical smart wheelchairs. Each wheelchair was equipped with two lidars, a 3D camera, odometers, user interfaces, and an embedded computer. During trials, users confirmed or rejected suggested routes via the wheelchair’s interface, allowing the system to learn and adapt to individual preferences and needs. The team’s findings suggest that AI-powered wheelchairs have the potential to significantly improve the lives of individuals with limited mobility, but challenges remain.
The Cost and Complexity of Intelligent Mobility
While the potential benefits of smart wheelchairs are clear, Pooja Viswanathan, CEO and founder of Toronto-based Braze Mobility, cautions that accessibility and practicality are crucial considerations. “Cost remains a major barrier,” Viswanathan says. “Funding systems are often not designed to support advanced add-on intelligence unless there is exceptionally clear evidence of value and safety.” Braze Mobility itself focuses on more incremental improvements, offering blind-spot sensors that can be added to existing wheelchairs to enhance safety.
Reliability is another key concern. Smart wheelchairs must function consistently in “the messy, variable conditions of daily life,” Viswanathan emphasizes. The “human factors” dimension – accounting for diverse cognitive, motor, sensory, and environmental needs – further complicates the development process. A one-size-fits-all solution is unlikely to be effective.
Louise Devinge, a biomedical research engineer at IRISA (Research Institute of Computer Science and Random Systems) in Rennes, France, highlights the technical challenges associated with increasing the complexity of these devices. “The more sensing, computation, and autonomy you add,” Devinge explains, “the harder it becomes to ensure robust performance across the full range of real-world environments that wheelchair users encounter.” Managing the flow of data from multiple sensors and ensuring seamless communication within the system is a significant undertaking.
A Partnership Between User and Technology
The DFKI researchers are addressing these challenges by focusing on creating a collaborative relationship between the user and the AI. The system isn’t intended to completely replace human control, but rather to provide intelligent assistance and enhance the user’s capabilities. This approach acknowledges the expertise and judgment of wheelchair users, recognizing that they often possess a remarkable ability to navigate their surroundings.
Mandel recalls his early work developing a head-joystick controlled wheelchair, and realizing that even individuals with severe disabilities demonstrated impressive navigational skills in confined spaces. “I realized…never underestimate what [wheelchair users] can do without it,” he says. This realization shaped his current research, emphasizing the importance of augmenting, rather than replacing, existing abilities.
Looking Ahead: Mainstream Adoption and Explainable AI
Mandel anticipates that smart wheelchairs could be available in the mainstream market within the next 10 years. However, Viswanathan believes that projects like REXASI-PRO, while currently beyond the reach of most consumers, are essential for long-term progress. “It reflects the more ambitious complete of the smart wheelchair spectrum,” she says, “and takes seriously the importance of trustworthy and explainable AI, which is essential in any mobility technology where safety, reliability, and user confidence are paramount.”
The DFKI researchers presented their work earlier this month at the CSUN Assistive Technology Conference in Anaheim, California. The ongoing development of these technologies, supported by organizations like the IEEE Foundation and fellowships like the Jon C. Taenzer fellowship, represents a significant step towards greater independence and quality of life for wheelchair users. The future of mobility isn’t just about building smarter machines; it’s about building machines that work *with* people, empowering them to navigate the world on their own terms.
The next step for the DFKI team involves refining the system’s algorithms and conducting further user testing to optimize performance and address remaining challenges. Continued research and development, coupled with a focus on affordability and accessibility, will be crucial to realizing the full potential of AI-powered wheelchairs.
What are your thoughts on the future of assistive technology? Share your comments below, and let’s continue the conversation.
