본문 바로가기

전체 글

[Python] DSRNN nav 코드 분석
Python 2022. 11. 30. [Python] DSRNN nav 코드 분석 Author: KangchanRoh Team: Reinforcement Learning Team @ CAI Lab Date: 2022/11/30 train.py train.py를 실행하면 main() 함수가 실행됨 Line 24. from crowd_nav.configs.config import Config에서 다음 항목에 관한 특성과 설정들을 다룰 수 있다 env reward function human robot noise orca, social force ppo SRNN training details Line 93. actor_critic 식별자로 아래 그림의 DSRNN 모델이 구현된 Policy 클래스 객체 생성 Line 99. rollouts 식별자로 학습 데이터 저장소 역할의 RolloutSto..
YOLO + Object Tracking
인공지능 환경세팅 2022. 11. 30. YOLO + Object Tracking Author: KangchanRoh Team: Reinforcement Learning Team @ CAI Lab Date: 2022/11/30 Benchmark COCO Benchmark (Real-Time Object Detection) Papers with Code - COCO Benchmark (Real-Time Object Detection) The current state-of-the-art on COCO is YOLOv7-E6E (36 fps). See a full comparison of 41 papers with code. paperswithcode.com MOT17 Benchmark (Multi-Object Tracking) Papers with Code - MOT17 Benchmark..
[Optimal Control] Paper Review: Cautious Model Predictive Control using Gaussian Process Regression
논문 리뷰 2022. 11. 28. [Optimal Control] Paper Review: Cautious Model Predictive Control using Gaussian Process Regression 0. Title - Cautious Model Predictive Control using Gaussian Process Regression 1. Authors - Lukas Hewing 2. Abstract - There is also an MPC approach that integrates a nominal system with an additive nonlinear part of the dynamics modeled as a Gaussian Process. - It 's a principled way of formulating the chance constrained MPC problem, which takes into account residual uncertainties provided by t..
[Optimal Control] Paper Review: Constrained Policy Optimization
논문 리뷰 2022. 11. 18. [Optimal Control] Paper Review: Constrained Policy Optimization 0. Title - Constrained Policy Optimization 1. Authors - Joshua Achiam 2. Abstract - systems that physically interact with or around humans should satisfy safety constraints. Recent advances in policy search algorithms have enabled new capabilities in highdimensional control, but do not consider the constrained setting. - We propose Constrained Policy Optimization (CPO), the first general-purpose..
[Optimal Control] Paper Review: Safe Exploration in Continuous Action Spaces
논문 리뷰 2022. 11. 17. [Optimal Control] Paper Review: Safe Exploration in Continuous Action Spaces 0. Title - Safe Exploration in Continuous Action Spaces 1. Authors - Joshua Achiam 2. Abstract - This paper is based on learning the policy in presence of hard constraints i.e the constraints should never be violated. In this work, they directly add to the policy a safety layer that analytically solves an action correction formulation per each state. This method largely builds upon the work of O..
[Optimal Control] Paper Review: A Review of Safe Reinforcement Learning: Methods,Theory and Applications
논문 리뷰 2022. 11. 17. [Optimal Control] Paper Review: A Review of Safe Reinforcement Learning: Methods,Theory and Applications 0. Title - A Review of Safe Reinforcement Learning: Methods, Theory and Applications 1. Authors - Shangding Gu 2. Abstract - When it comes to deploying RL in the real world, safety concerns are usually raised, leading to a growing demand for safe RL algorithms, such as in autonomous driving and robotics scenarios. 3. Motivation (1) Safety Policy. How can we perform policy optimization to search ..
[Optimal Control] Paper Review: A Comprehensive Survey on Safe Reinforcement Learning
논문 리뷰 2022. 11. 16. [Optimal Control] Paper Review: A Comprehensive Survey on Safe Reinforcement Learning 0. Title - A Comprehensive Survey on Safe Reinforcement Learning 1. Authors - Javier Garcia 2. Abstract - Safe RL의 정의: 어떠한 시스템의 성능을 보장하는 문제에서 return 값을 최대화하고, 학습하는 동안에 안정성을 위한 제약을 지키는 것 Definition of Safe RL: Maximizing the return in a problem that guarantees the performance of any system, and maintaining the constraints for stability during learning - Safe RL의 정의에 의한 두 가지 접근 1. 최적화 기준: 어떻게 하면 안..
[Optimal Control] Paper Review: Comparison of Deep Reinforcement Learning and Model Predictive Control for Adaptive Cruise Control
논문 리뷰 2022. 11. 7. [Optimal Control] Paper Review: Comparison of Deep Reinforcement Learning and Model Predictive Control for Adaptive Cruise Control 0. Title - Comparison of Deep Reinforcement Learning and Model Predictive Control for Adaptive Cruise Control 1. Authors - Yuan Lin 2. Abstract - 차량 추종 시나리오에서 심층 강화 학습(DRL)과 모델 예측 제어(MPC)를 비교, DRL은 DDPG 쓰고 MPC는 IPO 씀 Comparison of Deep Reinforcement Learning (DRL) and Model Predictive Control (MPC) in vehicle follow-up scenarios: DRL uses DDPG and MPC uses IPO - 결과는 모델링 오류가 없고 테스트 입력이 훈련 데이터 범위 ..
자율주행 논문 리뷰: Baidu Apollo EM Motion Planner
논문 리뷰 2022. 10. 4. 자율주행 논문 리뷰: Baidu Apollo EM Motion Planner 0. Title - Baidu Apollo EM Motion Planner (2018) 1. Authors - Haoyang Fan 2. Abstract - 안전성, 편안함, 확장가능성을 고려하여 Level 4 경로 계획을 다루는데 집중함 - 다차선 및 단차선 자율주행을 계층적으로 다룸 - 고속도로 및 저속 주행 시나리오 모두 확장 가능 3. Motivation - 기존의 방식의 문제점인 연산량이 많고 교통법규를 일반화하기 힘들며(차선의 종류마다 다른 교통 법규) 경로 안정성을 고려하지 못한다는 문제점을 해결하고자 함 - Non-convex 문제를 Convex 문제로 변환하여 최적의 솔루션을 찾고자 하였음 - 기존의 방식의 연산량 문제를 해결하고자 함 4. Contributions(Findings) - ..

티스토리툴바