Fuzzy Encoding-Decoding to Improve Spiking Q-Learning Performance in Autonomous Driving

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ghoreishee, Aref, Mishra, Abhishek, Zhou, Lifeng, Walsh, John, Das, Anup, Kandasamy, Nagarajan
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915943088652288
author Ghoreishee, Aref
Mishra, Abhishek
Zhou, Lifeng
Walsh, John
Das, Anup
Kandasamy, Nagarajan
author_facet Ghoreishee, Aref
Mishra, Abhishek
Zhou, Lifeng
Walsh, John
Das, Anup
Kandasamy, Nagarajan
contents This paper develops an end-to-end fuzzy encoder-decoder architecture for enhancing vision-based multi-modal deep spiking Q-networks in autonomous driving. The method addresses two core limitations of spiking reinforcement learning: information loss stemming from the conversion of dense visual inputs into sparse spike trains, and the limited representational capacity of spike-based value functions, which often yields weakly discriminative Q-value estimates. The encoder introduces trainable fuzzy membership functions to generate expressive, population-based spike representations, and the decoder uses a lightweight neural decoder to reconstruct continuous Q-values from spiking outputs. Experiments on the HighwayEnv benchmark show that the proposed architecture substantially improves decision-making accuracy and closes the performance gap between spiking and non-spiking multi-modal Q-networks. The results highlight the potential of this framework for efficient and real-time autonomous driving with spiking neural networks.
format Preprint
id arxiv_https___arxiv_org_abs_2604_16436
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Fuzzy Encoding-Decoding to Improve Spiking Q-Learning Performance in Autonomous Driving
Ghoreishee, Aref
Mishra, Abhishek
Zhou, Lifeng
Walsh, John
Das, Anup
Kandasamy, Nagarajan
Neural and Evolutionary Computing
Machine Learning
This paper develops an end-to-end fuzzy encoder-decoder architecture for enhancing vision-based multi-modal deep spiking Q-networks in autonomous driving. The method addresses two core limitations of spiking reinforcement learning: information loss stemming from the conversion of dense visual inputs into sparse spike trains, and the limited representational capacity of spike-based value functions, which often yields weakly discriminative Q-value estimates. The encoder introduces trainable fuzzy membership functions to generate expressive, population-based spike representations, and the decoder uses a lightweight neural decoder to reconstruct continuous Q-values from spiking outputs. Experiments on the HighwayEnv benchmark show that the proposed architecture substantially improves decision-making accuracy and closes the performance gap between spiking and non-spiking multi-modal Q-networks. The results highlight the potential of this framework for efficient and real-time autonomous driving with spiking neural networks.
title Fuzzy Encoding-Decoding to Improve Spiking Q-Learning Performance in Autonomous Driving
topic Neural and Evolutionary Computing
Machine Learning
url https://arxiv.org/abs/2604.16436