# Adaptive Optimization of Autonomous Vehicle Computational Resources for Performance and Energy Improvement

**Authors:** Saurabh Jambotkar, Longxiang Guo, Yunyi Jia  
**arXiv:** [2102.07000](https://arxiv.org/abs/2102.07000)  
**Date:** 2021-02-13  
**Summary page:** https://www.summarizepaper.com/en/paper/2102.07000v2/

## Key points

- Paper addresses challenge of computational power consumption in autonomous vehicles
- Proposes adaptive optimization method for resource allocation among vehicular subsystems
- Considers situational factors and different driving scenarios
- Aims to enhance performance and reduce energy consumption compared to existing arrangements
- Experiments validate effectiveness of the approach in improving performance and energy efficiency
- Contributes to addressing key challenge in autonomous vehicle development
- Offers potential improvements in both performance and energy efficiency
- Provides valuable insights into optimizing computational resource allocation for autonomous vehicles

## Summary

This paper titled "Adaptive Optimization of Autonomous Vehicle Computational Resources for Performance and Energy Improvement" addresses the challenge of computational power consumption in autonomous vehicles, particularly for tasks related to sensing and perception using artificial intelligence algorithms. To tackle this issue, the paper proposes an adaptive optimization method that dynamically allocates the onboard computational resources among multiple vehicular subsystems based on the specific contexts faced by the vehicle. By considering the situational factors, such as different autonomous driving scenarios, the proposed approach aims to enhance overall performance and reduce energy consumption compared to existing computational arrangements. The authors conducted experiments to validate their approach, designing various autonomous driving scenarios. The results demonstrated that their adaptive optimization method effectively improved both performance and energy efficiency in autonomous vehicles. This research contributes to addressing one of the key challenges in autonomous vehicle development by providing a solution to optimize computational resource allocation based on real-time contextual information. Overall, this paper presents a novel approach that tackles the issue of high computational power consumption in autonomous vehicles. By dynamically allocating resources based on contextual factors, it offers potential improvements in both performance and energy efficiency. The findings contribute valuable insights into optimizing computational resource allocation for autonomous vehicles and can guide future research in this domain.