Towards spiking analog hardware implementation of a trajectory interpolation mechanism for smooth closed-loop control of a spiking robot arm

TL;DR

This paper tackles the challenge of achieving smooth, closed-loop trajectory control for an event-based robotic arm using neuromorphic hardware. It introduces a two-network architecture consisting of a shifted Winner-Take-All (WTA) interpolator and a differential-position comparator, implemented on the DYNAP-SE2 and interfaced with the ED-Scorbot through the AER protocol. Experimental validation on a single joint demonstrates correct trajectory interpolation and accurate reach-detection, confirming hardware integration and real-time control potential. The work lays the groundwork for fully neuromorphic multi-joint robotic control and motivates future incorporation of neuromorphic memory and scaling to six degrees of freedom.

Abstract

Neuromorphic engineering aims to incorporate the computational principles found in animal brains, into modern technological systems. Following this approach, in this work we propose a closed-loop neuromorphic control system for an event-based robotic arm. The proposed system consists of a shifted Winner-Take-All spiking network for interpolating a reference trajectory and a spiking comparator network responsible for controlling the flow continuity of the trajectory, which is fed back to the actual position of the robot. The comparator model is based on a differential position comparison neural network, which governs the execution of the next trajectory points to close the control loop between both components of the system. To evaluate the system, we implemented and deployed the model on a mixed-signal analog-digital neuromorphic platform, the DYNAP-SE2, to facilitate integration and communication with the ED-Scorbot robotic arm platform. Experimental results on one joint of the robot validate the use of this architecture and pave the way for future neuro-inspired control of the entire robot.

Figures (7)

• Figure 1: Shifted WTA SNN architecture. Lines with arrow or circular ends represent excitation with AMPA or inhibition with SHUNT dendritic channels, respectively.
• Figure 2: SNN comparator architecture (a) at neural block level and (b) at computational neural units level.
• Figure 3: Diagram of the information flow and connection of the robotic arm and the DYNAP-SE2 platform.
• Figure 4: Robot controller and DYNAP-SE2 in loop-back.
• Figure 5: Raster plot of the spiking activity of the shifted SNN WTA during the sweep of all possible inputs for 4 possible reference positions for 1 joint with an offset of 1.