Learning a risk-aware policy is essential but rather challenging in unstructured robotic tasks. Safe reinforcement learning methods open up new possibilities to tackle this problem. However, the conservative policy updates make it intractable to achieve sufficient exploration and desirable performance in complex, sample-expensive environments. In this paper, we propose a dual-agent safe reinforcement learning strategy consisting of a baseline and a safe agent. Such a decoupled framework enables high flexibility, data efficiency and risk-awareness for RL-based control. Concretely, the baseline agent is responsible for maximizing rewards under standard RL settings. Thus, it is compatible with off-the-shelf training techniques of unconstrained optimization, exploration and exploitation. On the other hand, the safe agent mimics the baseline agent for policy improvement and learns to fulfill safety constraints via off-policy RL tuning. In contrast to training from scratch, safe policy correction requires significantly fewer interactions to obtain a near-optimal policy. The dual policies can be optimized synchronously via a shared replay buffer, or leveraging the pre-trained model or the non-learning-based controller as a fixed baseline agent. Experimental results show that our approach can learn feasible skills without prior knowledge as well as deriving risk-averse counterparts from pre-trained unsafe policies. The proposed method outperforms the state-of-the-art safe RL algorithms on difficult robot locomotion and manipulation tasks with respect to both safety constraint satisfaction and sample efficiency.

Safety comes first in many real-world applications involving autonomous agents. Despite a large number of reinforcement learning (RL) methods focusing on safety-critical tasks, there is still a lack of high-quality evaluation of those algorithms that adheres to safety constraints at each decision step under complex and unknown dynamics. In this paper, we revisit prior work in this scope from the perspective of state-wise safe RL and categorize them as projection-based, recovery-based, and optimization-based approaches, respectively. Furthermore, we propose Unrolling Safety Layer (USL), a joint method that combines safety optimization and safety projection. This novel technique explicitly enforces hard constraints via the deep unrolling architecture and enjoys structural advantages in navigating the trade-off between reward improvement and constraint satisfaction. To facilitate further research in this area, we reproduce related algorithms in a unified pipeline and incorporate them into SafeRL-Kit, a toolkit that provides off-the-shelf interfaces and evaluation utilities for safety-critical tasks. We then perform a comparative study of the involved algorithms on six benchmarks ranging from robotic control to autonomous driving. The empirical results provide an insight into their applicability and robustness in learning zero-cost-return policies without task-dependent handcrafting. The project page is available at https://sites.google.com/view/saferlkit.

The accurate detection and grasping of transparent objects are challenging but of significance to robots. Here, a visual-tactile fusion framework for transparent object grasping under complex backgrounds and variant light conditions is proposed, including the grasping position detection, tactile calibration, and visual-tactile fusion based classification. First, a multi-scene synthetic grasping dataset generation method with a Gaussian distribution based data annotation is proposed. Besides, a novel grasping network named TGCNN is proposed for grasping position detection, showing good results in both synthetic and real scenes. In tactile calibration, inspired by human grasping, a fully convolutional network based tactile feature extraction method and a central location based adaptive grasping strategy are designed, improving the success rate by 36.7% compared to direct grasping. Furthermore, a visual-tactile fusion method is proposed for transparent objects classification, which improves the classification accuracy by 34%. The proposed framework synergizes the advantages of vision and touch, and greatly improves the grasping efficiency of transparent objects.

在动态环境中，持续增强学习（CRL）的关键挑战是，随着环境在其生命周期的变化，同时最大程度地减少对学习的信息的灾难性忘记，随着环境在其一生中的变化而变化。为了应对这一挑战，在本文中，我们提出了Dacorl，即动态自动持续RL。 Dacorl使用渐进式上下文化学习了上下文条件条件的策略，该策略会逐步将动态环境中的一系列固定任务群集成一系列上下文，并选择一个可扩展的多头神经网络以近似策略。具体来说，我们定义了一组具有类似动力学的任务，并将上下文推理形式化为在线贝叶斯无限高斯混合物集群的过程，这些过程是在环境特征上，诉诸在线贝叶斯推断，以推断上下文的后端分布。在以前的中国餐厅流程的假设下，该技术可以将当前任务准确地分类为先前看到的上下文，或者根据需要实例化新的上下文，而无需依靠任何外部指标来提前向环境变化发出信号。此外，我们采用了可扩展的多头神经网络，其输出层与新实例化的上下文同步扩展，以及一个知识蒸馏正规化项来保留学习任务的性能。作为一个可以与各种深度RL算法结合使用的一般框架，Dacorl在稳定性，整体性能和概括能力方面具有一致的优势，而不是现有方法，这是通过对几种机器人导航和Mujoco Socomotion任务进行的广泛实验来验证的。

安全加强学习（RL）在对风险敏感的任务上取得了重大成功，并在自主驾驶方面也表现出了希望（AD）。考虑到这个社区的独特性，对于安全广告而言，仍然缺乏高效且可再现的基线。在本文中，我们将SAFERL-KIT释放到基准的安全RL方法，以实现倾向的任务。具体而言，SAFERL-KIT包含了针对零构成的侵略任务的几种最新算法，包括安全层，恢复RL，非政策Lagrangian方法和可行的Actor-Critic。除了现有方法外，我们还提出了一种名为精确惩罚优化（EPO）的新型一阶方法，并充分证明了其在安全AD中的能力。 SAFERL-KIT中的所有算法均在政策设置下实现（i），从而提高了样本效率并可以更好地利用过去的日志； （ii）具有统一的学习框架，为研究人员提供了现成的接口，以将其特定领域的知识纳入基本的安全RL方法中。最后，我们对上述算法进行了比较评估，并阐明了它们的安全自动驾驶功效。源代码可在\ href {https://github.com/zlr20/saferl_kit} {this https url}中获得。

安全的强化学习旨在学习最佳政策，同时满足安全限制，这在现实世界中至关重要。但是，当前的算法仍在为有效的政策更新而努力，并具有严格的约束满意度。在本文中，我们提出了受惩罚的近端政策优化（P3O），该政策优化（P3O）通过单一的最小化等效不受约束的问题来解决麻烦的受约束政策迭代。具体而言，P3O利用了简单的罚款功能来消除成本限制，并通过剪裁的替代目标消除了信任区域的约束。从理论上讲，我们用有限的惩罚因素证明了所提出的方法的精确性，并在对样品轨迹进行评估时提供了最坏情况分析，以实现近似误差。此外，我们将P3O扩展到更具挑战性的多构造和多代理方案，这些方案在以前的工作中所研究的情况较少。广泛的实验表明，在一组受限的机车任务上，P3O优于奖励改进和约束满意度的最先进算法。

在深入学习革命之前，许多感知算法基于运行时优化与强大的先前/正则化罚款。计算机视觉中的主要示例是光学和场景流。监督学习在很大程度上取代了明确规范化的必要性。相反，它们依靠大量标记的数据来捕获前面的统计数据，这并不总是随时可用的许多问题。虽然采用优化来学习神经网络，但是该网络的权重在运行时冻结。因此，这些学习解决方案是特定于域的，并不概括到其他统计上不同的场景。本文重新审视了依赖于运行时优化和强正规化的现场流动问题。这里的核心创新是在先前包含神经场景流，这利用神经网络的体系结构作为一种新型的隐式规范器。与基于学习的场景流方法不同，优化发生在运行时，并且我们的方法不需要脱机数据集 - 使其成为在自动驾驶等新环境中部署的理想选择。我们表明，专门在多层erceptrons（MLPS）上基于的架构可以用作现场流程。我们的方法持续竞争 - 如果没有更好的 - 结果在场景流基准上。此外，我们的神经先前的隐式和连续场景流量表示允许我们估计一系列点云序列的密集长期对应。密集运动信息由场景流场表示，其中通过积分运动向量可以通过时间传播点。我们通过累积激光雷达云序列来证明这种能力。

将钢筋学习（RL）扩展到推荐系统（RS）是有希望的，因为最大化RL代理的预期累积奖励达到了RS的目标，即提高客户的长期满意度。该目标的关键方法是离线RL，旨在从记录数据中学习政策。但是，商业RS中的高维操作空间和非平稳动态加剧了分配转移问题，这使得将离线RL方法应用于Rs是具有挑战性的。为了减轻从静态轨迹提取RL策略的动作分配转移问题，我们提出了基于不确定性的离线RL算法的价值惩罚Q学习（VPQ）。它通过不确定性的权重来惩罚回归目标中不稳定的Q值，而无需估计行为政策，适用于拥有大量项目的RS。我们从Q-功能合奏的方差中得出惩罚权重。为了减轻测试时间的分配转移问题，我们进一步介绍了评论家框架，以将拟议方法与经典RS模型相结合。在两个现实世界数据集上进行的广泛实验表明，该方法可以用作现有RS模型的增益插件。

Learning generalizable policies that can adapt to unseen environments remains challenging in visual Reinforcement Learning (RL). Existing approaches try to acquire a robust representation via diversifying the appearances of in-domain observations for better generalization. Limited by the specific observations of the environment, these methods ignore the possibility of exploring diverse real-world image datasets. In this paper, we investigate how a visual RL agent would benefit from the off-the-shelf visual representations. Surprisingly, we find that the early layers in an ImageNet pre-trained ResNet model could provide rather generalizable representations for visual RL. Hence, we propose Pre-trained Image Encoder for Generalizable visual reinforcement learning (PIE-G), a simple yet effective framework that can generalize to the unseen visual scenarios in a zero-shot manner. Extensive experiments are conducted on DMControl Generalization Benchmark, DMControl Manipulation Tasks, Drawer World, and CARLA to verify the effectiveness of PIE-G. Empirical evidence suggests PIE-G improves sample efficiency and significantly outperforms previous state-of-the-art methods in terms of generalization performance. In particular, PIE-G boasts a 55% generalization performance gain on average in the challenging video background setting. Project Page: https://sites.google.com/view/pie-g/home.

Humans can balance very well during walking, even when perturbed. But it seems difficult to achieve robust walking for bipedal robots. Here we describe the simplest balance controller that leads to robust walking for a linear inverted pendulum (LIP) model. The main idea is to use a linear function of the body velocity to determine the next foot placement, which we call linear foot placement control (LFPC). By using the Poincar\'e map, a balance criterion is derived, which shows that LFPC is stable when the velocity-feedback coefficient is located in a certain range. And that range is much bigger when stepping faster, which indicates "faster stepping, easier to balance". We show that various gaits can be generated by adjusting the controller parameters in LFPC. Particularly, a dead-beat controller is discovered that can lead to steady-state walking in just one step. The effectiveness of LFPC is verified through Matlab simulation as well as V-REP simulation for both 2D and 3D walking. The main feature of LFPC is its simplicity and inherent robustness, which may help us understand the essence of how to maintain balance in dynamic walking.