Offset-boosted reconfigurable multiscroll attractors in memristive Chua system for robot path planning

Existing memristive multiscroll systems mainly emphasize scroll-number expansion, while coordinated control of scroll position, spacing, clustering, and coexistence remains limited. This paper proposes a recursively composed memristor and integrates it into a classical Chua circuit to construct a four-dimensional chaotic system with configurable attractor geometry. By combining multilevel folding, sign-preserving operations, and offset controllers, the proposed design enables joint regulation of scroll number, position, spacing, and cluster organization, producing scroll merging, separation, and spatial reconfiguration. A microcontroller-based implementation further verifies the reproducibility and practical feasibility of the system. For path planning, representative coexisting attractor channels are organized into a coexisting-attractor bank (CAB). Their multichannel chaotic sequences are used to perturb Hybrid A* path-control points and generate diverse candidate paths. Experiments on three fixed grid maps show that, under the adopted evaluation settings, the complete CAB framework produces higher feasible-candidate ratios and candidate-path diversity than the selected baselines. These results provide a proof-of-concept demonstration of coexisting-attractor-derived multichannel signals for multi-candidate path generation.