ESTECH provides a wide range of technical support to achieve systems that are safe and comfortable for mobility users.
-
Development of coupled mechanical and control system models
-
Development of subjective evaluation indexes for comfortable autonomous driving
-
Construction of evaluation systems in virtual environments using driving simulators
-
Support for control logic development
-
Development of automatic tuning technology for control parameters
Emergency Evasive Control Model on Low-μ Roads
Accident prevention is one of the most important functional requirements for autonomous vehicles, and obstacle avoidance using steering control as well as speed control is a key technology. Especially for high-speed driving on low-μ roads, it is essential to balance the conflicting requirements of responsiveness and stability. We studied energy-optimal control that can achieve high vehicle stability through emergent steering in autonomous vehicles.
Automatic Parking Model Using Reinforcement Learning
Automatic parking identifies parking spaces using ultrasonic sensors and cameras, then automatically moves to a proper starting position and performs reverse parking.
This requires sufficient space to park in a single reverse maneuver. In narrow spaces, implementing complex algorithms that involve multiple switching operations—normally performed manually—requires situation-specific coding and increases implementation difficulty.
We realized an algorithm through reinforcement learning that enables parking from arbitrary positions and orientations with minimal control effort.
