Hardware-aware vs. Hardware-agnostic Energy Estimation for SNN in Space Applications

TL;DR

This paper compares hardware-aware and hardware-agnostic energy estimation methods for SNNs in space applications, revealing that actual energy savings depend on hardware type and data sparsity, emphasizing transparent evaluation.

Contribution

It introduces a comprehensive comparison of energy estimation methods for SNNs in satellite positioning, highlighting the importance of hardware considerations and data characteristics for accurate energy assessment.

Findings

Hardware-agnostic estimates overstate energy savings by 50-60%.

Significant energy savings occur only on neuromorphic hardware with high sparsity.

Data characteristics like dark pixel ratio greatly influence energy consumption.

Abstract

Spiking Neural Networks (SNNs), inspired by biological intelligence, have long been considered inherently energy-efficient, making them attractive for resource-constrained domains such as space applications. However, recent comparative studies with conventional Artificial Neural Networks (ANNs) have begun to question this reputation, especially for digital implementations. This work investigates SNNs for multi-output regression, specifically 3-D satellite position estimation from monocular images, and compares hardware-aware and hardware-agnostic energy estimation methods. The proposed SNN, trained using the membrane potential of the Leaky Integrate-and-Fire (LIF) neuron in the final layer, achieves comparable Mean Squared Error (MSE) to a reference Convolutional Neural Network (CNN) on a photorealistic satellite dataset. Energy analysis shows that while hardware-agnostic methods predict a consistent 50-60% energy advantage for SNNs over CNNs, hardware-aware analysis reveals that significant energy savings are realized only on neuromorphic hardware and with high input sparsity. The influence of dark pixel ratio on energy consumption is quantified, emphasizing the impact of data characteristics and hardware assumptions. These findings highlight the need for transparent evaluation methods and explicit disclosure of underlying assumptions to ensure fair comparisons of neural network energy efficiency.