MoDE-VLA study evaluates contact-rich manipulation with Sharpa Wave
An external research team evaluated IMCopilot and MoDE-VLA on a dual-hand Sharpa Wave configuration for gear assembly, charger insertion, tube rearrangement and apple peeling.
Evidence notes
- IMCopilot supplied reinforcement-learning skills for in-hand rotation during data collection and execution.
- MoDE-VLA separated visual, force and tactile processing and reported a 73% peel-completion ratio on repeated apple-peeling cycles.
Company context
Founded by leadership from Hesai Technology, with operations across Singapore (R&D), Shanghai (manufacturing), and Mountain View (business). Focuses on embodied AI systems combining dexterous manipulation hardware with model-driven control. Develops both robotic platforms and in-house AI models for long-horizon task execution.