Robotic grasping has been an active area of research since the dawn of robotics. With recent advancements in artificial intelligence, vision, planning and machine learning, robots are beginning to enter unstructured and unknown environments, such as warehouses.

Warehouse automation and, in particular, picking, is an increasingly popular application domain due to its significance in logistics operations. It usually involves a robot picking a list of items from a shelf, often in parallel with human workers. Different items are usually packed close together in bins and require an end-effector that is small and versatile. Due to the constrained and unstructured nature of the corresponding workspace, sometimes collisions with the environment are not easily avoidable.

This thesis outlines the design and fabrication of a hybrid end-effector that uses both suction and mechanical grasping courtesy of soft-robotics inspired fingers mounted on a flexible arm extension, both of which very robust and capable of sustaining multiple collisions without failure, which could otherwise be expensive and time-consuming. Furthermore, the combination of suction, grasping and soft robotics is a novel idea and offers additional benefits such as low weight and inexpensive components.