Robotic arm controlling based on a spiking neural circuit and synaptic plasticity. (January 2020)

• Record Type: Journal Article
• Authors: Wei, Hui; Bu, Yijie; Zhu, Ziyao
• Is Part Of: Biomedical signal processing and control. Volume 55(2020)
• Publication Date: 2020-01
• Subjects: Spiking neural circuit -- Robotic arm -- Synaptic plasticity -- Learning mechanism; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2019.101640
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: We simulate signal-processing mechanism of two types of neurons in cortex in detail to design a decision-making neural circuit. We apply forementioned neural circuit to control movement of a robot, and make it able to follow the position changing of a visual clue. This work is helpful to neurorobotics. Abstract: In many applications, it is a vital part to precisely control the movement of a robotic arm to reach a position. Currently the control of a robotic arm is achieved by solving complex kinematic equations. However, controlling the movements of limbs is a very basic ability for animals. Animals can accurately and quickly control their limbs to achieves goals. It is achieved by the inside biological neural mechanisms. Thus, a question worth researching arises naturally. That is can we control the movements of a robotic arm by mimicking animals' neural mechanisms? In our previous works, we have studied the biological neural mechanism under Drosophila's flight controlling behavior. In this paper, we applied the neural mechanism on the robotic arm controlling task, in which we need to control the robotic arm to reach or avoid a certain position according to the environment feedback. A spiking neural circuit is built to achieve the control of the robotic arm. And dopamine-modulated spike-timing-dependent plasticity is used to adjust the circuit connections according to the environment feedback. We show that with such a neural circuit and synaptic plasticity, the robotic arm can autonomously learn to reach a position or to avoid a position, given proper environment feedbacks.