Real-Time Neuromorphic Navigation: Guiding Physical Robots with Event-Based Sensing and Task-Specific Reconfigurable Autonomy Stack

TL;DR

The paper tackles real-time autonomous navigation on power-constrained robots by proposing a Reconfigurable Neuromorphic Navigation framework that fuses event-based perception (DVS and SNNs) with physics-informed planning and energy-aware control, designed to run on edge hardware such as the NVIDIA Jetson Nano. It presents a modular autonomy stack that is adaptable to ground and aerial platforms, and validates it with simulations and real-world demonstrations on TurtleBot and Bebop2, including obstacle-rich and moving-ring scenarios. Key contributions include end-to-end integration of perception, planning, and control within a neuromorphic, low-latency pipeline, physics-informed trajectory optimization, and comparative evaluations showing substantial improvements in maneuver time and energy efficiency over RGB-based baselines. The work demonstrates practical feasibility of neuromorphic navigation on resource-constrained hardware and outlines future directions for multi-robot collaboration, reinforcement learning extensions, and hardware-specific optimizations to further reduce latency and energy usage.

Abstract

Neuromorphic vision, inspired by biological neural systems, has recently gained significant attention for its potential in enhancing robotic autonomy. This paper presents a systematic exploration of a proposed Neuromorphic Navigation framework that uses event-based neuromorphic vision to enable efficient, real-time navigation in robotic systems. We discuss the core concepts of neuromorphic vision and navigation, highlighting their impact on improving robotic perception and decision-making. The proposed reconfigurable Neuromorphic Navigation framework adapts to the specific needs of both ground robots (Turtlebot) and aerial robots (Bebop2 quadrotor), addressing the task-specific design requirements (algorithms) for optimal performance across the autonomous navigation stack -- Perception, Planning, and Control. We demonstrate the versatility and the effectiveness of the framework through two case studies: a Turtlebot performing local replanning for real-time navigation and a Bebop2 quadrotor navigating through moving gates. Our work provides a scalable approach to task-specific, real-time robot autonomy leveraging neuromorphic systems, paving the way for energy-efficient autonomous navigation.

Figures (15)

• Figure 1: Comparison between biological neurons and spiking neuron models, highlighting dendrites, soma, axons, and the discrete spike timing (t) in Spiking Neural Networks (SNNs).
• Figure 2: Side-by-side comparison of conventional Artificial Neural Networks (ANNs) and Spiking Neural Networks (SNNs), showcasing continuous activation in ANNs versus discrete, event-driven spikes in SNNs.
• Figure 3: Comparison of conventional frame-based vision processing and event-based SNN processing, highlighting the low-power, high-temporal-resolution advantages of Dynamic Vision Sensors (DVS).
• Figure 4: left: Events are grouped spatially using a weighted neighborhood, and spiking neurons generate outputs for object detection and tracking. right: Input spikes are temporally filtered by spiking neurons; only dense event streams exceed the threshold, producing output spikes.
• Figure 5: Event-based Object Detection at Various Depths: Each sub-panel shows neuromorphic event output and a bounding box around a moving gate as it recedes from the camera (left to right). Although the event density decreases with increasing depth, the SNN isolates and tracks the gate in real time.