Dual-Brain Collaborative Navigation: Integrating Motor Imagery BCI With Brain Function-Inspired Trajectory Synthesis for Assistive Robotics

Brain–computer interface (BCI)-controlled navigation systems face a critical dilemma: continuous BCIs impose unsustainable cognitive load, while autonomous planners fail to adapt to human intent. We present a dual-brain collaborative framework with hierarchical cognitive separation: human operators provide sparse strategic commands (navigate to destination, move to target, return to start, step back) via motor imagery, while a brain-inspired robotic planner executes intelligent tactical navigation. The system integrates: First, asynchronous neural command interface achieving 94.73% offline accuracy and 89.57% online success with 97 ms latency; second, prefrontal trajectory synthesizer (PTS) demonstrating around 50% collision check reduction versus state-of-the-art algorithm across three motion planning benchmarks while maintaining >99% planning success. Hardware validation with 20 participants navigating physical mazes using Hiwonder TurboPi mobile robot achieved 84.5% task completion (169/200 trials), 42% cognitive load reduction (NASA Task Load Index (NASA-TLX), p< 0.001), and requiring 90.1% fewer commands than continuous BCI baselines. This mechatronic system establishes a paradigm for cognitively aligned assistive navigation through synergistic brain–machine collaboration.