.png){kind=icon}
.png){kind=icon}
Return
Toward a Dynamic Closed-Loop Neuromorphic Trajectory Interpolation Model: Hardware Emulation and Software Simulation
.png){kind=icon}
OA
AI
DOI：10.1109/TCSII.2025.3582362.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Neuromorphic engineering aims to incorporate the computational principles found in animal brains into modern technological systems. Controllers for robotic arms targeting specific positions often fall short in executing dynamically smooth trajectories, exhibiting rather segmented, step-like motions. This brief presents a closed-loop neuromorphic control system for event-based robotic arms, emphasizing a dynamic approach to trajectory interpolation through hardware emulation and software simulation. Our model employs a Shifted Winner-Take-All spiking network to interpolate reference trajectories and a spiking comparator network to ensure trajectory continuity against real-time positions, closing the control loop dynamically. The model’s implementation on various neuromorphic platforms highlights its flexibility and adaptability across distinct computational paradigms, such as analog hardware emulation and digital software simulation. Experimental results demonstrate the efficacy of the analog implementation in terms of robustness and energy efficiency, while the digital simulations produce precise and stable performance. These outcomes substantiate the model’s capacity to enhance robotic trajectory control and lay the foundation for the development of future neuromorphic robotic control systems.
Keywords:
Trajectory interpolation
spiking neural networks
DYNAP-SE2
SpiNNaker
neuromorphic control
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
I
IF:
0
Papers:
153
Citations:
0
View more organizations