SpikePingpong: Spike Vision-based Fast-Slow Pingpong Robot System

TL;DR

A Fast-Slow system architecture where System 1 provides rapid ball detection and preliminary trajectory prediction with millisecond-level responses, while System 2 employs spike-oriented neural calibration for precise hittable position corrections for strategic ball striking is developed.

Abstract

Learning to control high-speed objects in dynamic environments represents a fundamental challenge in robotics. Table tennis serves as an ideal testbed for advancing robotic capabilities in dynamic environments. This task presents two fundamental challenges: it requires a high-precision vision system capable of accurately predicting ball trajectories under complex dynamics, and it necessitates intelligent control strategies to ensure precise ball striking to target regions. High-speed object manipulation typically demands advanced visual perception hardware capable of capturing rapid motion with exceptional temporal resolution. Drawing inspiration from Kahneman's dual-system theory, where fast intuitive processing complements slower deliberate reasoning, there exists an opportunity to develop more robust perception architectures that can handle high-speed dynamics while maintaining accuracy. To this end, we present \textit{\textbf{SpikePingpong}}, a novel system that integrates spike-based vision with imitation learning for high-precision robotic table tennis. We develop a Fast-Slow system architecture where System 1 provides rapid ball detection and preliminary trajectory prediction with millisecond-level responses, while System 2 employs spike-oriented neural calibration for precise hittable position corrections. For strategic ball striking, we introduce Imitation-based Motion Planning And Control Technology, which learns optimal robotic arm striking policies through demonstration-based learning. Experimental results demonstrate that \textit{\textbf{SpikePingpong}} achieves a remarkable 92\% success rate for 30 cm accuracy zones and 70\% in the more challenging 20 cm precision targeting. This work demonstrates the potential of Fast-Slow architectures for advancing robotic capabilities in time-critical manipulation tasks.

Figures (7)

• Figure 1: Overview of SpikePingpong. The framework integrates two key stages: (1) Interception, using a Fast-Slow architecture for precise trajectory prediction, and (2) Striking, employing the IMPACT module to execute strategic returns via imitation learning. The system achieves a 92% overall success rate and 70% in high-precision targeting tasks.
• Figure 2: Framework of SpikePingpong. The system comprises two integrated components: (1) A Fast-Slow perception architecture, where System 1 delivers rapid trajectory prediction using RGB-D data, while System 2 functions as a Spike-Oriented Neural Improvement Calibrator to refine the estimated hittable position; and (2) The IMPACT module, which facilitates strategic motion planning and control, enabling tactical return placement via imitation learning.
• Figure 3: System Overview. Our system integrates three key subsystems: (1) a coordinate system for spatial tracking and transformation, (2) a multi-frequency control system with Fast-Slow system, IMPACT, and an EGM controller, and (3) a robot system based on the ABB IRB-120 arm equipped with a standard table tennis racket.
• Figure 4: Visualization of ball-racket contact precision with and without System 2. Spike-based camera images show the ball center (green) and racket center (red) at contact moment. The reduced offset with the Fast-Slow System demonstrates improved ball interception accuracy.
• Figure 5: Comparison of ball capture qualit. Conventional RGB camera with motion blur at 60 fps, Spike camera with crisp imagery at 20,000 fps, demonstrating the advantage of ultra-high frame rate for fast-moving object detection.