The field of robotic manipulation is moving towards more complex and dynamic tasks, such as manipulating deformable objects, handling occlusions, and interacting with uncertain environments. Researchers are exploring innovative approaches, including reinforcement learning, imitation learning, and sensor-space based control, to enable robust and efficient manipulation. These advances have the potential to significantly improve the capabilities of robotic systems in various applications, including manufacturing, healthcare, and service robotics. Notable works in this area include the development of hierarchical reinforcement learning frameworks for occluded grasping, linear control strategies for heterogeneous object manipulation on soft surfaces, and sensor-space based kinematic control for redundant soft manipulators. Some particularly noteworthy papers are:

• A study on leveraging extrinsic dexterity for occluded grasping, which proposes a hierarchical reinforcement learning framework for robust grasping in complex environments.
• A paper on heterogeneous object manipulation, which introduces a simple and robust PID-based linear control strategy for manipulating diverse objects on a soft robotic surface.