Can robotic swarms navigate using learning rules devised for brain dynamics?

The lab of Professor Kechen Zhang of the Department of Biomedical Engineering (BME) and researchers from the JHU/Applied Physics Laboratory, including Kavli NDI steering committee member Grace Hwang, have a new paper in the journal Biological Cybernetics detailing a new theory relating the neural dynamics of memory in the brain to the autonomous control of robotic swarms. 

The paper's lead author, Research Associate Joseph Monaco in BME, showed that neural oscillatory coding theories and Hebbian forms of synaptic plasticity could be transformed to emulate the mathematics of the motion of many robotic agents acting in concert. In particular, the collective behaviors necessary for an artificial swarm to autonomously navigate fragmented environments with randomly distributed rewards can emerge from the same computations thought to be performed by hippocampal circuits in navigating rats. The resulting framework of 'cognitive swarming' provides a bridge between seemingly disparate disciplines that may in fact share underlying computational principles. 

Share this page onFacebookTwitter