PyRep: Bringing V-REP to Deep Robot Learning

TL;DR

The paper introduces PyRep, a Python-friendly, speed-optimized variant of V-REP designed to accelerate data collection for robot-learning experiments. It implements three core modifications: a direct Python-controlled simulation loop for drastically reduced latency, a modern OpenGL 3.0+ renderer with shadows, and a high-level PyRep API for easy scene and robot manipulation. The authors report substantial speedups—up to 10^4×—over the original Python remote API, enabling rapid prototyping of reinforcement learning, imitation learning, state estimation, and computer vision workflows. PyRep aims to bridge the gap between powerful robotics simulators and data-hungry learning methods by providing practical tooling for large-scale simulation data collection.

Abstract

PyRep is a toolkit for robot learning research, built on top of the virtual robotics experimentation platform (V-REP). Through a series of modifications and additions, we have created a tailored version of V-REP built with robot learning in mind. The new PyRep toolkit offers three improvements: (1) a simple and flexible API for robot control and scene manipulation, (2) a new rendering engine, and (3) speed boosts upwards of 10,000x in comparison to the previous Python Remote API. With these improvements, we believe PyRep is the ideal toolkit to facilitate rapid prototyping of learning algorithms in the areas of reinforcement learning, imitation learning, state estimation, mapping, and computer vision.

Figures (2)

• Figure 1: Example images of environments using the new renderer.
• Figure 2: PyRep API Example. Many more examples can be seen on the GitHub page.