1 00:00:00,140 --> 00:00:03,000 This is a grocery packing warehouse in the south-east of London. 2 00:00:03,000 --> 00:00:04,620 Folks have ordered their groceries online, 3 00:00:04,620 --> 00:00:06,599 and instead of people walking up and down the aisles 4 00:00:06,599 --> 00:00:09,119 of a traditional grocery store to pick up items, 5 00:00:09,119 --> 00:00:13,560 these little devices move more than a million items every day between them. 6 00:00:13,560 --> 00:00:16,760 Instead of trying to build machines that fit into a world designed for humans, 7 00:00:16,760 --> 00:00:20,179 this entire facility, the size of seven football fields, 8 00:00:20,179 --> 00:00:22,679 is custom-built to make this as efficient as possible. 9 00:00:22,679 --> 00:00:29,739 I cannot get across on camera just how incredibly large this whole place is. 10 00:00:29,739 --> 00:00:34,119 The top speed of those boxes is about 14 kilometres an hour and they pass within five millimetres 11 00:00:34,119 --> 00:00:35,600 of each other when they move. 12 00:00:35,600 --> 00:00:39,979 So how many robots are there in the hive? 13 00:00:39,979 --> 00:00:42,880 Well that depends on how you count them. 14 00:00:42,880 --> 00:00:45,340 What you're looking at behind me is the hive. 15 00:00:45,340 --> 00:00:48,320 We've got the grid which is the metal structure. 16 00:00:48,320 --> 00:00:53,119 It's filled with products so that product is housed within what we call totes or these 17 00:00:53,119 --> 00:00:56,060 open boxes that you can see at the top of the grid. 18 00:00:56,060 --> 00:01:01,240 We have around 2,300 bots and they move the stock around the warehouse so that we can 19 00:01:01,240 --> 00:01:04,159 pair it up with customer orders to be delivered. 20 00:01:04,159 --> 00:01:08,840 The bots will move in the X and Y direction, they can also move in the Z direction. 21 00:01:08,840 --> 00:01:13,819 So we have a cable hoist system that means that they can lower down a gripper assembly 22 00:01:13,819 --> 00:01:16,920 and grab onto the stock that we have in the grid. 23 00:01:16,920 --> 00:01:21,560 In some areas we have as deep as 21 totes all in a line. 24 00:01:21,560 --> 00:01:26,120 We have sensors underneath the bot that mean that every time it passes over one of the 25 00:01:26,120 --> 00:01:31,000 squares on the grid, it has a laser underneath it that can tell when it moves over that cell. 26 00:01:31,000 --> 00:01:34,879 So it will say, I've passed a cell, I've passed a cell, I've passed a cell. 27 00:01:34,879 --> 00:01:39,579 We have around 58,000 different types of individual products within the grid. 28 00:01:39,579 --> 00:01:43,980 We'll use the huge amounts of data that we gather to understand what customers are most 29 00:01:43,980 --> 00:01:48,180 likely to order and place them in the easiest to reach locations for bots. 30 00:01:48,180 --> 00:01:52,739 All of that combined means that we can have stock into the warehouse and out again from 31 00:01:52,739 --> 00:01:56,400 our suppliers into a delivery van in just five hours. 32 00:01:56,400 --> 00:01:58,939 The robots deliver their crates over to the packing area. 33 00:01:58,939 --> 00:02:02,819 Some of that packing is done by humans, but there are mechanical arms as well. 34 00:02:02,819 --> 00:02:06,459 One of the big problems they've got to solve is that grocery packaging is designed for 35 00:02:06,459 --> 00:02:11,400 people and the interactions between all the objects they deal with can be complicated. 36 00:02:11,400 --> 00:02:14,780 Currently we're standing underneath the grid structure that we saw earlier. 37 00:02:14,780 --> 00:02:18,580 The robots that we saw are actually dropping off products to the station. 38 00:02:18,860 --> 00:02:20,699 Once they're dropped off, they become in control of the robot, 39 00:02:20,780 --> 00:02:25,740 and the robot decides how to pick them before the grid robot picks up the tote and takes it away again. 40 00:02:25,960 --> 00:02:31,159 There's two 3D cameras above the totes, so it takes an image of the storage box before it picks from it, 41 00:02:31,400 --> 00:02:33,939 decides on the best grasp point, confirms that the product is correct, 42 00:02:34,219 --> 00:02:38,319 and then places it into the customer shopping, and it's doing all of that on live customer orders. 43 00:02:38,800 --> 00:02:42,060 We make sure that we don't mix something that was heavy with something that's soft. 44 00:02:42,060 --> 00:02:47,039 There are things that a robot can't pack, things such as very, very heavy, very large or very fragile items, 45 00:02:47,199 --> 00:02:48,699 but there's a lot of range in between that. 46 00:02:49,039 --> 00:02:52,379 It's fairly safe to call each individual packing arm a robot. 47 00:02:52,680 --> 00:02:55,960 Each one has a separate controller and it has to make a lot of decisions in real time 48 00:02:55,960 --> 00:02:58,419 using cameras and sensors on each individual arm. 49 00:02:58,840 --> 00:03:01,680 But here in the Hive, it's not quite as simple. 50 00:03:02,259 --> 00:03:05,560 Controlling all of this is what we refer to as the Hive Mind, 51 00:03:05,740 --> 00:03:09,400 so this artificial intelligence system that controls everything. 52 00:03:09,400 --> 00:03:16,400 Once the stop comes in, from that exact point that it enters the system, the hive mind will track its every single movement. 53 00:03:16,400 --> 00:03:28,400 The hive mind will tell the bots exactly where they need to stop, but it is the bot's control system that allows it to follow a movement profile and stop exactly as the hive mind expects it to. 54 00:03:28,400 --> 00:03:35,400 There's just a 5mm gap between bots, so that's both as they pass each other and if they stop next to a bot as well. 55 00:03:35,400 --> 00:03:41,300 This bot with the orange light may have experienced something within its sensors and all of those readings it's taking 56 00:03:41,300 --> 00:03:47,000 that when it feeds that back to the hive mind it's not quite what the hive mind was expecting 57 00:03:47,000 --> 00:03:50,159 so it just brings it to a safe stop for investigation. 58 00:03:50,800 --> 00:03:55,120 We do have a number of grid operators that sit at the edge of the grid 59 00:03:55,120 --> 00:03:59,219 and they're able to use our CCTV system that we have above the grid 60 00:03:59,219 --> 00:04:02,020 to look at exactly what's going on with that bot 61 00:04:02,020 --> 00:04:06,120 and they can try and fix that remotely and send commands to the bot wirelessly. 62 00:04:06,759 --> 00:04:09,560 The boundary between individual and group is a little blurry. 63 00:04:09,740 --> 00:04:13,860 Is a termite colony and a collection of individuals working together or one big superorganism? 64 00:04:14,020 --> 00:04:16,500 Is the microbiome in your gut part of you? 65 00:04:16,779 --> 00:04:19,019 Because in the last couple of decades, scientists have started to work out 66 00:04:19,019 --> 00:04:23,060 whether bacteria in your digestive system produce and consume neurotransmitters 67 00:04:23,060 --> 00:04:24,699 that affect how you think. 68 00:04:25,199 --> 00:04:28,740 But most people wouldn't consider their gut bacteria to be part of them. 69 00:04:28,740 --> 00:04:33,240 We tend to ascribe intelligence and identity to things that act at roughly our scale. 70 00:04:33,240 --> 00:04:35,980 It's easy to look at the hive, see the moving boxes, and go, 71 00:04:35,980 --> 00:04:37,980 OK, 2,000 robots. 72 00:04:37,980 --> 00:04:41,740 But with only one controller, and the boxes just following the paths they're given, 73 00:04:41,740 --> 00:04:44,980 I'd say there's a good argument that that is one robot 74 00:04:44,980 --> 00:04:47,980 with 2,000 ways to interact with the world. 75 00:04:47,980 --> 00:04:50,980 Thank you so much to everyone here at Ocado. 76 00:04:50,980 --> 00:04:53,980 There is a link to them and more about their technology in the description.