The 2020 World Robot Contest-BCI Controlled Robot Contest was held from December 3rd to 6th in Foshan, Guangdong Province. The event was co-hosted by Chinese Institute of Electronics and Tsinghua University - under the guidance of National Natural Science Foundation of China. Students from School of Artificial Intelligence and Automation distinguished themselves in R&D Group.
Two teams led by Professor Dongrui Wu took part in the finals (WRCF) and championships (WRCC) respectively. The first, composed of 6 graduate students, ranked second in ERP (Event related potential) without Training Data and third in ERP with Training Data, reaping the first prize in R&D Group. The other group, consisting of 2 graduates and 4 undergraduates, entered the championships of Emotional BCIs and Motor Imagery Without Training, winning third prize in R&D Group. It’s the second time for Professor Wu’s teams to achieve a placing in the contest after winning the first place in the R&D Group in 2019.
BCI (Brain-Computer Interface) technology is to collect, recognize and transform electrical activities - and characteristic signals of the nervous system - in order to transmit instructions from human brains directly to a specified terminal, thus making it easier to operate a robot. It is of innovative significance in human-robot communication, and has been widely applied to fields including rehabilitation and recreation, making great contributions to improving the lives of people with disabilities.
This year’s contest offers 7 problems based on ERP, Motor Imagery, SSVEP (Steady-State Visual Evoked Potentials) and Emotional BCIs. Professor Wu’s teams performed with distinction in ERP, which is a kind of special brain-evoked potential that can reflect the neurophysiological changes of the brain during cognitive processes. In ERP with Training Data, the model needs to be established according to the calibration data collected by each user. While in ERP without Training Data, there’s no data given, which makes it more challenging and significant. Guided by Professor Wu, the two teams overcame challenges using their self-developed transfer learning algorithm, eventually winning the third and second place respectively in the online contest of real-time data collection.
The World Robot Contest has been successfully held 6 times since 2015, attracting over 150,000 participants from more than 20 countries around the world. Widely praised as the “Olympics of Robotics” by the international mainstream media, WRC is now a professional event in robotics with a profound influence at home and abroad. BCI Controlled Robot Contest displays the pioneering achievements of robots in numerous areas, and arranges specific tasks around the annual “hot technologies”, aiming to promote core strategic breakthroughs in global robotics.
Written by: Peishan Li
Edited by: Rong He, Scott, Yumeng Peng
Source: School of Artificial Intelligence and Automation