自動運転車に「自己」は必要か：ホメオスタシスを応用した理論モデル
The need of a self for self-driving cars a theoretical model applying homeostasis to self driving
自動運転車に内部状態を監視するホメオスタシス的アーキテクチャを導入し、外環境を自己状態のずれとして解釈することで、自律性・安全性・効率を高める理論モデルを提案した。
分類: cs.HC, cs.RO
原文アブストラクト
This paper explores the concept of creating a "self" for self-driving cars through a homeostatic architecture designed to enhance their autonomy, safety, and efficiency. The proposed system integrates inward focused sensors to monitor the car's internal state, such as the condition of its metal bodywork, wheels, engine, and battery, establishing a baseline homeostatic state representing optimal functionality. Outward facing sensors, like cameras and LIDAR, are then interpreted via their impact on the car's homeostatic state by quantifying deviations from homeostasis. This contrasts with the approach of trying to make cars "see" reality in a similar way to humans and identify elements in their reality in the same way humans. Virtual environments would be leveraged to accelerate training. Additionally, cars are programmed to communicate and share experiences via blockchain technology, learning from each other's mistakes while maintaining individualized training models. A dedicated language for self-driving cars is proposed to enable nuanced interpretation and response to environmental data. This architecture allows self-driving cars to dynamically adjust their behavior based on internal and external feedback, promoting cooperation and continuous improvement. The study concludes by discussing the broader implications for AI development, potential real-world applications, and future research directions.