Brain–computer interfaces (BCIs) have gained popularity in recent years. Among noninvasive BCIs, EEG-based systems stand out as the primary approach, utilizing the motor imagery (MI) paradigm to discern movement intentions. Initially, BCIs were predominantly focused on nonembedded systems. However, there is now a growing momentum towards shifting computation to the edge, offering advantages such as enhanced privacy, reduced transmission bandwidth, and real-time responsiveness. Despite this trend, achieving the desired target remains a work in progress. To illustrate the feasibility of this shift and quantify the potential benefits, this paper presents a comparison of deploying a CNN for MI classification across different computing platforms, namely, CPU-, embedded GPU-, and FPGA-based. For our case study, we utilized data from 29 participants included in a dataset acquired using an EEG cap for training the models. The FPGA solution emerged as the most efficient in terms of the power consumption–inference time product. Specifically, it delivers an impressive reduction of up to 89% in power consumption compared to the CPU and 71% compared to the GPU and up to a 98% reduction in memory footprint for model inference, albeit at the cost of a 39% increase in inference time compared to the GPU. Both the embedded GPU and FPGA outperform the CPU in terms of inference time.

Self-propelled wheelchairs are challenging to drive on off-road routes or require enormous physical effort in situations where the gradient exceeds 8%. For most people, these situations are too strenuous and therefore impractical. This work presents an innovative plug-and-play system to electrify a manual wheelchair. With the constraint of not irreversibly modifying the original wheelchair, a mechanical system is developed to allow motors to turn the rear wheels by friction. Torque and power requirements have been derived by modeling and simulating in Matlab the physical system. To manufacture a working prototype, a control system using standard components is devised, and object-oriented firmware is developed with the C++ programming language. The result is a plug-and-play portable kit that can be adapted to wheelchairs of any size, and the expandable Human-Machine-Interface (HMI) makes it ready to expand the audience that can use it.

Driving a motorized wheelchair is not without risk and requires high cognitive effort to obtain good environmental perception. Therefore, people with severe disabilities are at risk, potentially lowering their social engagement, and thus, affecting their overall well-being. Therefore, we designed a cooperative driving system for obstacle avoidance based on a trained reinforcement learning (RL) algorithm. The system takes the desired direction and speed from the user via a joystick and the obstacle distribution from a LiDAR placed in front of the wheelchair. Considering both inputs, the system outputs a pair of forward and rotational speeds that ensure obstacle avoidance while being as close as possible to the user commands. We validated it through simulations and compared it with a vector field histogram (VFH). The preliminary results show that the RL algorithm does not disruptively alter the user intention, reduces the number of collisions, and provides better door passages than a VFH; furthermore, it can be integrated on an embedded device. However, it still suffers from higher jerkiness.

Social integration is an essential part of the life of every human being, but for people with disabilities, there are many situations in which it is still very low. For instance, in sports and outdoor gaming, there is a barrier between players with and without disabilities. Individuals with disabilities play sports almost exclusively with disabled players, not only during official events such as Paralympic games but also in occasional sports groups, while the mixing of people with and without disabilities in sports activities is a key factor of social inclusion. In order to allow a person with motor-skill impairments to play on the same ground as their non-disabled peers, we developed a novel piece of Assistive Technology that lets a person with motor-skill impairments to control a system acting as a goalkeeper during a non-professional football match, with approximately the same performances as a goalkeeper without motor-skill impairments. This electro-mechanical system is composed of a three-meter-long metal guideline and a human-shaped dummy sliding along it. The system is equipped with a high-torque battery-powered direct-current motor and it is controlled by means of electronic boards and sensors to ensure safety and good usability also for players with severe mobility impairments. The results of a pilot testing demonstrated the robustness and high degree of usability of the system, enabling people with motor-skill impairments to competitively participate in matches with non-disabled peers.

Operating a powered wheelchair involves significant risks and requires considerable cognitive effort to maintain effective awareness of the surrounding environment. Therefore, people with significant disabilities are at a higher risk, leading to a decrease in their social interactions, which can impact their overall health and well-being. Thus, we propose an intelligent driving-assistance system that innovatively uses Transformers, typically employed in Natural Language Processing, for navigation and a retrieval mechanism, allowing users to specify their destinations using natural language. The system records the areas visited and enables users to pinpoint these locations through descriptions, which will be considered later in the retrieval phase. Taking a foundational model, the system is fine-tuned with simulated data. The preliminary results demonstrate the system’s effectiveness compared to non-assisted solutions and its readiness for deployment on edge devices.