A Strategy Adaptive Adjustment Deep Reinforcement Learning Method with Behavior Cloning for Mobile Robot Navigation
Bibliographic record
Abstract
Deep reinforcement learning plays a key role in the navigation of mobile robots in various environments. However, current reinforcement learning training methods face challenges that suffer from: 1) the randomness of the initial parameters slows down the training speed during the training phase, and 2) the potential conflict between goal-directed tasks and obstacle avoidance tasks leads to difficulties in action selection during the navigation phase. To address these issues, we propose the Strat-egy Adaptive Reinforcement Learning(SARL) method, which combines imitation learning during the training phase, adjusts learning strategies dynamically based on agent feedback, and dynamically generates way points to guide the robot during the navigation phase. Specifically, we propose a method that dynamically integrates Twin Delayed Deep Deterministic Policy Gradient (TD3) and Behavior Cloning (BC), using dynamic feedback from the agent to adjust the data and prioritize training on high-quality samples, thereby resolving the issue of slow training speed resulting from the randomness of initial parameters. Additionally, we propose a composite constraint reward function that considers both the angle and distance to the target point, while incorporating a parking penalty, thereby effectively addressing the issue of reward sparsity. Finally, we design an intelligent replanning mechanism that generates path points based on obstacle density when the robot is stuck, resolving conflicts between goal-directed tasks and obstacle avoidance during navigation. The experimental results across multiple scenarios demonstrate that during the training process, the SARL method outperforms the state-of-the-art methods in terms of training efficiency and navigation success rate. Furthermore, SARL provides new insights for achieving efficient, flexible, and reliable autonomous navigation systems, offering improvements and enhancement potential for robot navigation in practical application scenarios.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".