WEBVTT 00:00.000 --> 00:06.480 We are actually aiming for a new type of automation that breaks the 00:06.480 --> 00:08.280 limits of the factory. 00:10.660 --> 00:16.320 Nowadays, robots are primarily automated in the factory. 00:16.960 --> 00:20.440 Typical manoeuvrability tasks are performed on the machines. 00:21.720 --> 00:27.020 Or is it about placing the parts quickly in short periods of time? 00:29.520 --> 00:36.640 What we see and think about is automation that leaves these limits of 00:36.640 --> 00:38.700 the factory. 00:39.260 --> 00:41.580 For example, when it comes to assistance tasks. 00:41.580 --> 00:49.360 Assistance tasks mean that the worker is supported, for example, by 00:49.360 --> 00:50.400 robot functions. 00:51.000 --> 00:53.280 He works hand-in-hand with the device. 00:54.000 --> 01:00.760 That means, for example, a worker sets up a process. 01:01.340 --> 01:05.240 That means, for example, the screw-in is prepared, but the actual 01:05.240 --> 01:07.200 procedure is carried out by the robot at the end. 01:07.200 --> 01:09.620 So this is an assistance. 01:10.120 --> 01:13.660 The person saves time, but does the demanding things first at the 01:13.660 --> 01:13.980 beginning. 01:15.220 --> 01:18.660 And that's exactly how you can imagine that the robot performs 01:18.660 --> 01:27.020 assistance tasks when performing or testing any workpieces or tools, 01:28.320 --> 01:34.560 which have to be done regularly with high precision and where human 01:34.560 --> 01:36.160 errors should be eliminated as much as possible. 01:37.120 --> 01:39.840 And these are the ideas that we pursue here. 01:40.400 --> 01:45.300 I see you also have a grip attachment in your hand. 01:45.340 --> 01:45.620 Exactly. 01:45.920 --> 01:46.400 Is that part of it? 01:47.360 --> 01:49.120 That is well recognized. 01:49.240 --> 01:49.880 It goes with it. 01:50.360 --> 01:52.840 This is a grip hand. 01:53.500 --> 01:57.560 In contrast to the grippers you see here on the robot, they are 01:57.560 --> 02:01.260 characterized by the function of parallel fingers. 02:02.840 --> 02:06.560 That means, two fingers drive parallel to each other and grab the 02:06.560 --> 02:06.960 objects. 02:07.180 --> 02:10.300 This is the typical task setting as it occurs in the industry. 02:11.580 --> 02:17.320 But if I want to have assistance tasks or, for example, want to use 02:17.320 --> 02:22.060 the robot in the human environment, let's say in the hospital, in the 02:22.060 --> 02:27.600 nursing home, at home, then of course the robot also has to cope with 02:27.600 --> 02:29.460 the conditions accordingly. 02:29.880 --> 02:32.820 For example, it has to be able to grab door handles, window handles, 02:32.980 --> 02:34.840 drawers and such things. 02:36.240 --> 02:42.280 And it can't do that with an ordinary industrial gripper, but the grip 02:42.280 --> 02:44.700 function has to be a bit more complicated. 02:45.260 --> 02:48.420 That's why we're thinking about how the grip can be for the future. 02:48.860 --> 02:50.220 I'll hold it in front of the camera. 02:50.220 --> 02:54.120 As you can see, the hand consists of three fingers. 02:54.640 --> 02:56.960 Each of these fingers has two joints. 02:57.180 --> 02:58.840 So similar to how it is with us humans. 02:59.480 --> 03:01.500 The joints can be controlled individually. 03:02.460 --> 03:07.020 In addition, it is possible that these two fingers oppose each other. 03:07.020 --> 03:08.920 That means they are twisted against each other. 03:09.460 --> 03:14.280 Then you also have the possibility to grab precise parts here. 03:14.280 --> 03:18.660 Otherwise, it is also possible to grab large parts, cylindrical or 03:18.660 --> 03:21.140 cubical parts with this hand. 03:22.260 --> 03:27.360 It fits directly on the robot via this interface here and can be 03:27.360 --> 03:28.220 operated there. 03:29.060 --> 03:33.700 The special feature of this development is that it is equipped with 03:33.700 --> 03:37.540 tactile sensor surfaces. 03:38.270 --> 03:46.240 This means that these surfaces can capture the pressure in a matrix in 03:46.240 --> 03:48.660 several places, which is then practiced. 03:48.900 --> 03:55.600 This allows you to detect contours of the object, workpiece or door 03:55.600 --> 03:56.360 handle. 03:56.680 --> 04:00.980 And you can also determine whether it was grabbed optimally. 04:01.900 --> 04:04.180 So it's very similar to how we humans do it. 04:04.180 --> 04:09.000 We first look at an object to be grasped with our eyes, then we grasp 04:09.000 --> 04:09.580 the position. 04:09.960 --> 04:14.520 But when it comes to really grasping the object, we use our tactile 04:14.520 --> 04:15.600 skills for this. 04:16.180 --> 04:20.480 For this, we need our five-finger hand. 04:20.560 --> 04:23.820 But we have found that three fingers would be sufficient for what the 04:23.820 --> 04:24.560 robot needs. 04:26.680 --> 04:29.240 This development has, by the way, been carried out together with the 04:29.240 --> 04:29.680 KIT. 04:29.680 --> 04:37.460 So we cooperated there and made this great development together.