PhD: Brain Controlled Wheelchair

This was my PhD Project, from January 2004 to December 2007. It was conducted at the National University of Singapore, under the supervision of Etienne Burdet, Marcelo Ang and Christian Laugier.

The resulting thesis is available for download here (PDF 3.4MB).

Abstract:

The Brain Controlled Wheelchair (BCW) is a simple robotic system designed for people, such as locked-in people, who are not able to use physical interfaces like joysticks or buttons. Our goal is to develop a system usable in hospitals and homes with minimal infrastructure modifications, which can help these people regain some mobility. The main challenge is to provide continuous and precise 2D control of the wheelchair from a Brain Computer Interface, which is typically characterized by a a very low information transfer rate. Besides, as design constraints, we want our BCW to be safe, ergonomic and relatively low cost.

The strategy we propose relies on 1) constraining the motion of the wheelchair along predefined guiding paths, and 2) a slow but accurate P300 EEG brain interface to select the destination in a menu. This strategy reduces control to the selection of the appropriate destination, thus requires little concentration effort from the user. Besides, the trajectory is predictable, which contributes to reduce stress, and eliminates frustration that may be associated with trajectories generated by an artificial agent. Two fast BCIs are proposed to allow stopping the wheelchair while in motion. A hybrid BCI was developed to combine the slow P300 BCI used for destination selection with a faster modality to stop the wheelchair while in motion.

Experiments with healthy users were conducted to evaluate performances of the BCIs. We found that after a short calibration phase, the destination selection BCI allowed the choice of a destination within 15 seconds on average, with an error rate below 1%. The faster BCI used for stopping the wheelchair allowed a stop command to be issued within 5 seconds on average. Moreover, we investigated whether performance in the STOP interface would be affected during motion, and found no alteration relative to the static performance. Finally, the overall strategy was evaluated and compared to other brain controlled wheelchair projects. Despite the overhead required to select the destination on the interface, our wheelchair is faster than others (36% faster than MAIA): thanks to the motion guidance strategy, the wheelchair always follows the shortest path and a greater speed is possible. Comparison was also performed using a cost function that takes into account traveling time and concentration effort; our strategy yields by far the smallest cost (the best other score is 72% larger).

This work resulted in a novel brain controlled wheelchair working prototype. It allows to navigate in a familiar indoor environment within a reasonable time. Emphasis was put on user's safety and comfort: the motion guidance strategy ensures smooth, safe and predictable navigation, while mental effort and fatigue are minimized by reducing control to destination selection.

Pictures:

Principle of a BCI: signal from the brain are acquired, digitized and analyzed to extract commands that can be used to control a computer or a device. The brain signal signal is usually EEG signals due to the relative low cost and ease of use of the hardware, but other technologies are possible, such as fNIR. The signal processing part can take various forms, but the most successful applications resort to statistical machine learning techniques. Applications include spelling, computer mouse control, and prosthesis or robot control. In the case of wheelchair control, special care must be given to user's safety: a wrong command can put the user in a dangerous situation.

The BCI visual display (this is an animated image, if you cannot see the animation click on the image or download it to view it locally on your computer). Items are flashed one by one in a random order. To select one item the user focuses his or her attention on it. Around 300ms after the target is presented, a positive potential peak appears in the EEG signal (the P300 signal). The interface displays a list of predefined locations. Upon selection of a location the wheelchair automatically starts moving toward it following the appropriate guiding path.

Publications

• B. Rebsamen, E. Burdet, C. Guan, H. Zhang, C.L. Teo, Q. Zeng, M. Ang and C. Laugier, Controlling a Wheelchair Indoors using Thought, IEEE Intelligent Systems, March/April 2007.

• Q. Zeng, C.L. Teo, B. Rebsamen, and E. Burdet, A Collaborative Wheelchair System, IEEE Trans. on Neural Systems and Rehabilitation Engineering (TNSRE), Vol. 16, No. 2, 2008.

• Q. Zeng, E. Burdet, B. Rebsamen, and C.L. Teo, Collaborative Path Planning for a Robotic Wheelchair, Disability and Rehabilitation: Assistive Technology, 2008

• B. Rebsamen, C. Guan, H. Zhang, C. Wang, C.L. Teo, M. Ang, E. Burdet, A Brain Controlled Wheelchair to Navigate in Familiar Environments, IEEE Trans. on Neural Systems and Rehabilitation Engineering (TNSRE), Vol. 18, No. 6, 2010.

• B. Rebsamen, E. Burdet, C. Guan, H. Zhang, C.L. Teo, Q. Zeng, M. Ang and C. Laugier, Controlling a wheelchair using a BCI with low information transfer rate, IEEE Int. Conf. on Rehabilitation Robotics (ICORR), 2007

• Q. Zeng, C.L. Teo, B. Rebsamen, and E. Burdet, Evaluation of the Collaborative Wheelchair Assistant System, in IEEE Int. Conf. on Rehabilitation Robotics (ICORR), 2007.

• Q. Zeng, E. Burdet, B. Rebsamen, and C.L. Teo, Experiments on Collaborative Learning with a Robotic Wheelchair, in Int. Convention for Rehabilitation Engineering and Assistive Technology (i-CREATe), 2007.

• B. Rebsamen, E. Burdet, C. Guan, H. Zhang, C.L. Teo, Q. Zeng, M. Ang and C. Laugier, A Brain-Controlled Wheelchair Based on P300 and Path Guidance, IEEE Int. Conf. on Biomedical Robotics and Biomechatronics (BioRob), 2006.

• Q. Zeng, B. Rebsamen, E. Burdet and C.L. Teo, Design of a Collaborative Wheelchair with Path Guidance Assistance, IEEE Int. Conf. on Robotics and Automation (ICRA), 2006.

• B. Long, B. Rebsamen, E. Burdet and T.C. Leong, Development of an elastic path controller, IEEE Int. Conf. on Robotics and Automation (ICRA), 2006, pp 493-498.

• B. Long, B. Rebsamen, E. Burdet, T.C. Leong and H.Y. Yu, Elastic Path Controller for Assistive Devices, Proc. 27th Int. Conf. of IEEE-EMBS, 2005.