Describing it as the world’s first integration of a quantum-inspired optimization computer directly inside an autonomous mobile robot- the demonstration conducted in Kawasaki and Nisshin, Japan, marks a shift from server-based processing to on-device decision-making for autonomous navigation.
The system uses Toshiba’s Simulated Bifurcation Machine, a quantum-inspired optimization platform running on conventional hardware rather than requiring dedicated quantum infrastructure. Unlike true quantum computers, SBM operates on standard semiconductors including FPGAs, GPUs, and ASICs, making the technology practical for embedded applications.
The core challenge addressed is multi-object tracking in dynamic environments where people and vehicles frequently cross paths or become temporarily obscured. Traditional tracking systems often lose objects or swap identities when overlaps occur. Toshiba developed a multi-object tracking algorithm implemented on an embedded FPGA using proprietary circuit design techniques, enabling the robot to maintain accurate tracking even during occlusions.
In field demonstrations, the autonomous robot combined SBM-based tracking with positional analysis to dynamically adjust predicted object occupancy zones. This approach reduced unnecessary avoidance manoeuvres while maintaining safe navigation through crowded spaces containing both static and moving obstacles.
The companies plan to expand embedded quantum-inspired optimization to transport robots in factories, autonomous construction and agricultural machinery, infrastructure monitoring, smart cities, and energy management systems. The technology addresses growing demand for autonomous systems in logistics and smart mobility amid labor shortages.
