Figure 1 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
CAD rendering of the experimental setup. The red cone represents the proximity sensor field of view.
Source publication
The gripper is the far end of a robotic arm. It is responsible for the contacts between the robot itself and all the items present in a work space, or even in a social space. Therefore, to provide grippers with intelligent behaviors is fundamental, especially when the robot has to interact with human beings. As shown in this article, we built an in...
Contexts in source publication
Context 1
... prototype was actuated by means of two external pressure regulators, and was equipped with customdesigned fingers and with a six-axis, force/torque (F/T) sensor which served to measure the grip force and to retrieve the CoP components. An IR-based proximity sensor (PS) with conic field of view was mounted on a support which was in turn located close to the gripper, as illustrated in the rendering of Figure 1. This sensor was utilized to detect the object presence and its positioning before grasping it. ...
Context 2
... that the maximum full-scale error of the employed F/T sensor is ±10 N for F Z , resulting in very high theoretical errors (several cms) at low forces/torques as the ones in Figure 9. Concerning the PS sensor, its calibration for both offset and crosstalk was carried out according to the manufacturer indications. To this end, a specific setup ( Figure 10) was constructed; the PS was mounted on its support as in Figure 4, and located in front of a target surface. The support was in turn inclined in such a way that the PS was parallel to the target surface. ...
Context 3
... achieve higher frequencies with good precision, it is reasonable to set both parameters so that TB = I MP − γ, where γ is as little as possible, usually 5 ms. From Figure 11, when TB and I MP are too short, i.e., TB = 15 ms and I MP = 20 ms, the SD of the PS measurement is too high (top subplot). This leads to low accuracy and introduces measurement uncertainty. ...
Context 4
... we discuss the full experiments with the tested objects. To analyze different grasping conditions due to different geometry and hardness, Figures 13 and 14 illustrate some representative trials with the holed metal piece and with the soft cylinder, respectively. A linear slip was induced to the former, whereas the latter was disturbed with a rotational slip. ...
Context 5
... e.g., Figure 13: until 1 s, both grasping force F Z and the CoP are null, i.e., the gripper is in IDLE state. This can be seen in the first two subplots. ...
Context 6
... fingers close as the desired force F D reaches the first level F 1 = 7.5 N, typical of the OBJECT DETECTED transition. Figure 13. Representative trial on the holed metal piece, linear slip. ...
Context 7
... of the above considerations apply to the case of the soft cylinder in Figure 14. For this object, as well as for the styrofoam piece, a lower desired force F 2 was set, i.e., F 2 = 20 N. In contrast to Figure 13, this time the CoPN value is rather higher than the pos one as long as the object is stably grasped. ...
Context 8
... of the above considerations apply to the case of the soft cylinder in Figure 14. For this object, as well as for the styrofoam piece, a lower desired force F 2 was set, i.e., F 2 = 20 N. In contrast to Figure 13, this time the CoPN value is rather higher than the pos one as long as the object is stably grasped. The gripper can manage the object grasping regardless of the CoP location and of the actual object distance measured by the PS (provided that this one falls below its threshold). ...
Citations
... Our system combines RGB-D cameras, closed-loop control servo motors, and custom soft tactile sensors to determine an optimal packing order for unknown grocery items in a safe, dynamic and online fashion. grippers through using vision for tactile and force sensing [12]- [14] or incorporating rigid sensing elements within soft systems [15], [16], there has been relatively little exploration into applying soft robots that use contemporary sensor fusion techniques to complex online applications. ...
... The most popular combination of sensor modalities is in the use of vision for tactile sensing, where the high resolution of a camera or time-of-flight sensor is used to track the deformation of a soft surface to get tactile information [12], [13]. Others incorporate rigid elements to provide proprioception within their soft structure, occasionally supplementing this with further tactile sensors [15], [16]. We build on this approach and our previous work [33] by choosing a strategy where we incorporate proprioceptive feedback from the rigid servo motors that drive our soft gripper with the soft tactile sensors and external vision system for our multimodal approach. ...
Robots are helping humans perform tasks, especially in industry. The collaboration between humans and machines is considered to be among the principal factors underlying the so called
Industry 4.0
. To boost such a form of collaboration, robots could be provided with a higher degree of instrumentation, which undoubtedly resorts to a reliable sensorization. In this regard, the robot end-effector is central, being often the interaction organ of the robot. The present article aims at delivering a device, namely instrumented finger (IF), that can be mounted on robotic end-effectors such as grippers. The device is able to sense the grasping force and the associated bending moment thanks to an integrated load cell. It has also a tactile matrix through which the center of pressure can be determined, as well as the slip. Further, the device electronics includes an inertial sensor. Cables are protected by covers which hide the inner metallic body and the electronics. The device is compact, robust and precise, and constitute a promising solution for the instrumentation of robotic grippers. In the present article, the IF functionalities, such as measurement of force, torque, acceleration and center of pressure, will be demonstrated. Further, it will be shown the IF integration on a real industrial gripper mounted on a robotic arm, executing slip compensation experiments and mass estimation of grasped items.
