Grasping multiple object types (versatile object grasping) with a single gripper is always a challenging task in robotic manipulation. Different types of grippers, including rigid and soft, have been developed to try to achieve the task. However, each gripper type is still restricted to specific object types. In nature, many insects can be observed to use only one tarsus mechanism to cope with several tasks. They have a very high grasping capability with objects and can adhere to a variety of surface types. Inspired by insect tarsus, this paper proposes a novel underactuated, single cable-driven, compliant gripper design. The structure of the gripper is based on the hornet tarsus morphology with a proportional scale. An additional pulley-like structure is introduced to increase the generated grasping torque. To maintain the ability to automatically rebound back to the original position, a torsion spring is implemented at each joint. In order to stably grasp and hold objects, soft adhesive pads with an asymmetric sawtooth-like surface structure are attached at the tarsus segments. The performance of this insect tarsusinspired gripper with the soft pads is evaluated by grasping 35 different objects of various sizes, shapes, and weights for comparison with industrial soft and rigid grippers. The proposed gripper shows a 100% success rate in grasping all objects, while the soft and rigid gripper success rates are 81.90% and 91.43% on average, respectively. We finally demonstrate the use of our gripper installed on a robot arm for pick-and-place and pouring tasks.

No abstract

Passive and active hybrid pipeline inspection gauges (PIGs) have been used for in-pipe inspection. While a passive PIG cannot control its speed, the hybrid version can achieve this by using an integrated valve specifically designed and embedded in the PIG. This study proposes a generic new method for speed adaptation in PIGs (called MC-PIG) by introducing a generic, modular, controllable, external valve unit add-on for attaching to existing conventional (passive) PIGs with minimal change. The MC-PIG method is based on the principle of morphological computation with closed-loop control. It is achieved by regulating/computing the PIG's morphology (i.e., a modular rotary valve unit add-on) to control bypass flow. Adjustment of the valve angle can affect the flow rate passing through the PIG, resulting in speed regulation ability. We use numerical simulation with computational fluid dynamics (CFD) to investigate and analyze the speed of a simulated PIG with the valve unit adjusted by proportional-integral (PI) control under various in-pipe pressure conditions. Our simulation experiments are performed under different operating conditions in three pipe sizes (16″, 18″, and 22″ in diameter) to manifest the speed adaptation of the PIG with the modular valve unit add-on and PI control. Our results show that the PIG can effectively perform real-time adaptation (i.e., adjusting its valve angle) to maintain the desired speed. The valve design can be adjusted from 5 degrees (closed valve, resulting in high moving speed) to a maximum of 45 degrees (fully open valve, resulting in low moving speed). The speed of the PIG can be regulated from 0.59 m/s to 3.88 m/s in a 16″ pipe at 4.38 m/s (in-pipe fluid velocity), 2500 kPa (operating pressure), and 62 °C (operating temperature). Finally, the MC-PIG method is validated using a 3D-printed prototype in a 6″ pipe. Through the investigation, we observed that two factors influence speed adaptation; the pressure drop coefficient and friction of the PIG and pipeline. In conclusion, the results from the simulation and prototype show close characteristics with an acceptable error.