Brain-inspired Edge Robotics: From Neuromorphic Sensing to Autonomous Systems

As artificial intelligence continues to drive advances in computing hardware, the rapid increase in model size and computational complexity is placing ever-growing demands on hardware resources. This trend is particularly challenging for small robotic platforms, such as aerial robots, legged robots, and compact ground robots. These systems operate under tight constraints on size, weight, power consumption, computational resources, and communication bandwidth, while requiring low-latency perception, information processing, decision-making, and control. Conventional AI pipelines can impose substantial computational and energy costs, motivating alternative approaches to enable efficient intelligence directly at the robotic edge. Brain-inspired artificial intelligence and novel neuromorphic architectures offer promising solutions by exploiting sparse, event-driven, temporal, and distributed forms of computation.Recent advances demonstrate the potential of combining specialized neuromorphic sensing, and spiking neural networks with autonomous robotic systems. Event-based vision provides asynchronous, high-temporal-resolution sensing that is particularly well suited to dynamic and resource-constrained platforms. Neuromorphic processors have enabled increasingly sophisticated robotic applications, including autonomous drone flight using a fully neuromorphic vision-to-control pipeline, energy-efficient obstacle avoidance for aerial robots, and on-device localization on a hexapod robot using an integrated event-based sensor, spiking neural network, and neuromorphic processor. Neuromorphic approaches are also being explored for computationally demanding robotic control, including model predictive control on quadrupedal platforms. These developments highlight the emerging opportunity to move beyond isolated neuromorphic components toward tightly integrated, autonomous robotic systems in which sensing, learning, perception, planning, and control are performed efficiently at the edge.This Research Topic aims to bring together researchers working at the intersection of neurorobotics, neuromorphic computing, neuroscience, embodied intelligence and autonomous robotics. We particularly encourage contributions addressing small-scale and resource-constrained robotic platforms, where energy efficiency, latency, adaptability and computational autonomy are critical. Contributions may address neuromorphic and brain-inspired methods implemented in dedicated hardware, conventional edge-AI hardware or hybrid architectures, provided that they incorporate a meaningful neuromorphic, event-driven or biologically inspired component.Topics of interest include, but are not limited to: • Neuromorphic approaches for autonomous robotics • Embodied and neurorobotic intelligence • Spiking neural networks for robotic perception, learning, planning and control • Brain inspired navigation, locomotion, and decision making for edge and resource constrained robotics • On-device learning, continual learning and adaptation • Energy- and latency-efficient intelligence for resource-constrained robots • Event-based and neuromorphic sensing for robotics • Sensor fusion combining event-based, conventional and other asynchronous sensors • Real-world deployment, benchmarking and evaluation of brain-inspired robotic systemsBy bringing together perspectives from neuroscience, robotics, artificial intelligence and neuromorphic engineering, this Research Topic seeks to advance the development of efficient, adaptive, and autonomous robotic systems capable of operating with intelligence directly at the edge. We particularly welcome contributions that demonstrate brain-inspired and neuromorphic approaches in real-world robotic systems and investigate when and why these approaches can provide advantages over conventional computing paradigms. Topic editors: Alper Yegenoglu, Florian Röhrbein, Guido Trensch Source: https://www.frontiersin.org/research-topics/85510/brain-inspired-edge-robotics-from-neuromorphic-sensing-to-autonomous-systems