Robots can now find their way in blocked or dangerous places without GPS. Using ideas from insects, birds, and rodents, they navigate smarter and safer.
Robots can transform how critical tasks such as search and rescue operations, infrastructure inspections, and complex maintenance jobs are performed autonomously. One major challenge for these systems is navigating complex and unfamiliar environments without constant human guidance. Traditional reliance on Global Positioning Systems (GPS) becomes problematic in obstructed areas, including underground locations or zones affected by disasters.
Researchers at the Beijing Institute of Technology are driving a major advancement with a GPS-independent navigation framework inspired by nature’s most skilled navigators: insects, birds, and rodents. This innovation leverages nature’s time-tested navigation strategies to push the boundaries of robotic autonomy.
Conventional autonomous robots rely heavily on GPS for positional awareness. However, GPS fails in environments where signals cannot reach, such as dense forests, tunnels, or disaster sites with heavy debris.
A Bio-Inspired Three-Part Framework
- Insect-Inspired Path Integrator: Modeled on insects’ natural navigation, this component uses a spiking neural network to track movements in real time. Acting like an internal pedometer, it improves positional accuracy.
- Bird-Inspired Multisensory Fusion: Inspired by migratory birds, this module combines data from multiple sensors using a Bayesian filter. It maintains a steady heading even when some sensors fail.
- Rodent-Inspired Cognitive Mapping: Based on the rodent hippocampus, this system builds cognitive maps updated with key landmarks. It optimizes energy use and focuses computational resources on the most relevant information.
Together, these components form a neuromorphic framework for robust navigation without relying on GPS.
Performance and Field Testing
Real-world tests show significant improvements such as positional drift dropped by 41%, energy efficiency rose by 60%, and recovery from sensor failures improved by 83%. A fault-tolerant mechanism called “degeneracy” ensures other components can compensate if one fails.
This navigation system has broad potential in disaster response, planetary exploration, and undersea missions where GPS and traditional sensors are unreliable. Future versions may incorporate adaptive learning to further mimic the flexibility of living systems.
