Logistics, in general, is one of the most common application areas for mobile robots. Actually, the revolution of mobile robots began from AGVs (Automated Guided Vehicles) more than 50 years ago in warehouses and factories. They were simple vehicles towing trailers while following some fixed tracks or wires on the floor. We have come a long way from those times and nowadays the most advanced AGVs are full-blooded service robots and they come in many forms and sizes and the goods are not only towed but often carried onboard the robot. All of them share the same requirements, as all the other autonomous mobile robots: they need to know their positions, they need to have situational awareness, they need to be able to fulfill their mission objectives (normally this means moving safely and effectively something from point A to point B) and they need to be able to work as a piece of a larger system, including not only the robots, but the human operators, other mobile machines, and the warehouse/factory as a whole. Localisation of the machines can be based on many different solutions. Most common one includes rotating laser+reflectors, but in many applications some sort of tags on the floor are widely used as well. Both have one similar weakness, and that is the need for supportive infrastructure, whether it is a grid of stickers on the floor or a set of reflectors on the walls and on the ceiling. The system will not work without those and any major modifications in the warehouse usually requires some level of extra work regarding the infrastructure updates. Another widely used, infra-structure free, solution is based on industry-grade 2D LiDARs onboard the machine. The LiDARs will create a 2D map of the environment and then localize themselves on that map. To be useful, the system must obviously have a way to validate the map continuously and update it if some larger more permanent changes have been detected. This solution is working beautifully, but it has one problem which is naturally due to the 2D nature of the used LiDARs. As a result, these machines do not have the needed 3D situational awareness around them. To compensate for that, you basically have two options. You either add some additional sensors for the situational awareness (extra LiDARs, cameras, ultrasonic sensors, etc.) or you make sure that the working environment is so deterministic that 2D understanding, together with some emergency ultrasonic sensors and bumpers, is good enough. In plain English this means, that you are controlling very precisely what is happening in the warehouse, and you prevent/reduce any unplanned activities, like human operators wandering around the working areas or placing pallets to temporary locations, especially without notifying the Fleet Management System (FMS) and thus the higher warehouse level control system. To avoid the above problems (i.e., the need to have extra infrastructure, the rigidness of the system, the need to have additional onboard sensors and the limitations for the operation areas), GIM Robotics is offering an alternative solution. Our GIM-Locator, GIM-Mapper and GIM-Observer are basically all built on the use of 3D LiDARs. LiDARs are used to create and update a 3D model of the environment, to localize the robot in that environment and to provide a precise 3D situational awareness 360 degrees around the machine.
Last mile delivery
It is a widely known fact, that the last mile (or kilometer) delivery before the parcel/package reaches the end destination causes a large portion of the total transportation costs. There are many reasons for that, but the core reason is naturally the economics of scale. Transporting thousands/millions parcels with an ocean-going ship is so much more cost-effective than to have a single parcel transported from a local post office to a person living in an apartment on the fifth floor. To reduce that high cost per delivered parcel, the study for autonomous last-mile delivery robots has been extremely active for, at least, the past 20 years. There are two main paths: the airborne and the land-based solutions. As we have stayed out of the drone business, our focus has been on the land-based solutions. Although there are some great solutions using modern walking machines for the task, the majority rely on wheel-based solutions. We have been active on this field, basically from day one, and during the last two years we have been participating in the LMAD consortium (later company) run activities, first in France (Nokia Bell Labs Campus) with our first robot and later in several pilots in Finland with the next generation machine. The third generation last-mile delivery robot is just about to start the first proper outdoor testing in Espoo. Just like with any other autonomous mobile robot application, the last mile delivery case requires localization, situational awareness and task/mission performing capabilities. GIM-Locator localizes the machine in any weather conditions with centimeter-level accuracy. GIM-Mapper creates and updates the 3D map of the environment and GIM-Observer provides the needed real-time situational awareness around the robot. All those modules have been tested extensively. Our pilot in France has been running nearly continuously for almost two years.
Terminal logistics means here the logistics actions, which are taking place in locations, which are devoted for loading and unloading cargo. The definition is naturally very generic and includes everything from unloading a truck to one of the platforms outside a central warehouse serving grocery stores to full-scale operations of a fully automated harbor. Regardless of the volume, type, and location of those terminals, they all need the same functionalities. Trucks, tractors, carriers and so on, must be able to localize themselves, make and update the map of the environment and understand what is happening around them. In addition, they must work together with other relevant stakeholders (machines, robots, humans) under the supervision of the overall Mission Control, while performing those actual logistics operations, such as transporting new containers from a ship to their intermediate storing locations or moving frozen goods from the truck to a specific cold room. As the term terminal logistics is so wide, it is almost impossible to provide any detailed description of how the job should be done. It is evident that the large facilities have invested tons of money to secure the safe and productive performance of the automated vehicles mostly with some external infrastructure including tags, cameras, special RF solutions etc. And once again, we state that our GIM-Locator localizes the machine in any weather condition with centimeter-level accuracy, our GIM-Mapper creates and updates the 3D map of the environment, and our GIM-Observer provides the needed real-time situational awareness around the robot. And those do that without any need for external infrastructure. Should those systems be readily available, our Offerings will naturally use them to secure a higher level of redundancy.
Shops in general have been void of any types of robots until last decade. Main reason has been the lack of good enough use cases in respect to the maturity level of the relevant research field. Unless operated solely during the times when doors are closed for the public, the machines must be able to work safely among humans, sometimes even in crowded situations. The most common tasks these machines have been performing are transporting some goods from an unloading place (ref. to Terminal logistics) to storages and then from storages to the actual shop for manual placing to shelves. Only recently, two other use cases have emerged and got a lot of attention and funding. First one is the cleaning of the floors, while customers are present in the shop, and the other is analyzing the shelves of the shop for inventory (e.g., the number of items / product) and quality reasons (e.g., the price tags are correct). Cleaning requires, besides the obvious cleaning operation performing platform, constant interactions with the customers, while still performing the scheduled cleaning tasks. That, on the other hand, requires beside the normal localization (GIM-Locator) and environmental modeling (GIM-Mapper), some real-time situational awareness capabilities (GIM-Observer). Shop inventory application requires the same functionalities and on top of that, it also requires a sophisticated AI-flavored camera-based tool for the actual visual inspection. That part we leave for our partners. GIM Robotics has been, and still is, very much involved in these retail related activities with our proven commercial partners. These activities will be further strengthened in a brand-new Business Finland funded and VTT led Co-Innovation Project called “Multi-purpose Service Robotics as Operator Business” (MURO).