#15 Navigation Technologies for AGVs/AMRs
Matias do you know that there is one navigation technology which has not changed since the beginnings of agvs I don't know by heart I just know it's super confusing and every time I talk to potential mobile robot customers they are all confused about those different Technologies that's why we thought it might be a good idea to have one episode just dedicated to the navigation Technologies and to hopefully create some more understanding in the community automation Awakenings your weekly dose of best practices for Logistics automation welcome back to the next episode of the automation Awakenings podcast we are happy to have you on board with us and as we already said the topic of navigation Technologies is a really confusing one so let's try to shed some light into this uh this topic and go through those different Technologies maybe also discuss a little bit the pros the cons and also the the advantages that they have yes that sounds uh very interesting and let's start with um the differentiation between localization and navigation here we already dive into a confusing topic and um suppliers tend to mix up this two words so for sure an AG needs both right yeah it has to navigate and it has to locate itself within the environment yeah and I think you can distinguish those two functions with with very easy questions so localization answers the question where am I as a mobile robot on the other hand navigation would answer the question how do I get to my target or my destination and depending on which technology is answering the question you could say it's navigating or localization and in some cases to confuse you guys even more this technology can do both but maybe let's take it step by step and let's start with those Technologies who are only there to do the navigation yes I think these two questions uh will help us in this episode a lot and let's start very easy the Magnetic Tape yeah yeah it's there since the 1950s and from technology point of view this type of navigation technology has not changed yeah so we are still using it in current AGS even yeah they are following in a very simple way a magnetic tape on the floor so it's basically glued to the floor right Ollie yes it's either glued to the floor or drilled into the floor oh my very old type of applications okay yeah good um actually I have a fun fact about the Magnetic Tape um so when I prepared for this episode I found out that the Magnetic Tape is still the most commonly used navigation technology for mobile robots out there so no way Matias no that's that seems to be the case um so yes yeah it shows that it's still valid it's doing the job and um it obviously has its advantages which and that's the same for the other let's say basic um navigation Technologies like for example line guidance or induction uh tape or induction uh cable that the accuracy is extremely high yeah so I read that um the tolerance of inaccuracy is around 2 mm yeah sure that's awesome and it's not very costly of course we have a initial effort um during the implementation but then um it's reliable and you can use the the HV with the magnetic tape for a very long time and I know several applications um which last since yeah many many years even 10 years or more and they operate still in a in a very reliable way so they are yeah good old dinosaurs um which yeah no they don't have to be replaced if it's they're working yeah and I would even say that there is one more falling into this category and it's data metric so QR codes glued to the floor right yeah I would say this is um already The Next Step because at least you don't have to glue a uh a tape but every 1 to 1.5 M usually you have a unique QR code and here is a big difference to the Magnetic Tape because the Magnetic Tape is stupid right but the QR code already is unique M so the robot will know where it is so here we come again to localization and navigation so after reading the QR code on the floor you will know your exact position on the map right so maybe let's cluster these ones as non laser based navigation Technologies right can we say that yeah absolutely cool you don't need any laser for that you just need to read um the barcode or you need to read detect the Magnetic Tape yeah no laser required so in which kind of mobile robots app applications are these non- laser based Technologies used yeah you can always use them um you can use them everywhere uh that's the easy answer because they uh are working with um many environments they have no issues with um light yeah with sunlight for example or even in dark environments uh they will work humidity all of these yeah environmental factors um they don't play a role for this kind of Technology they are super reliable and stable in all kind of environments cool perfect so maybe let's you uh let's move into the the next category uh which is the laser based navigation Technologies and on my list as let's say the next Evolution step I have the simple laser reflector navigation technology so to say yes uh exactly here we already go one step further we need to add reflectors yeah um no QR codes on the floor but reflectors on fixed positions which are then also fixed on the map okay yeah that's important so we have this one time effort and um you need to teach um these reflectors very exact into this map so the robot will rely on this fixed positions and you must never ever change the position of these reflectors so if you hit one of them uh with a forklift or whatever um yeah you you need to repair it um immediately okay it might even happen that your whole mobile robots operation will stop only because one reflector is not there any more seriously yes because usually the robots need three of these reflectors to know where they are at the same time so the whole time they need to be in touch with three different reflectors yeah that's why it's also called triangular um location localization and if one is missing it only sees two positions and how can it know where it is with only two lines you know if you imagine as a a line being uh like yeah on the map yeah um you need the third line to have your exact position okay understood so that sounds a little bit more prone to environmental effects than Magnetic Tape or um say QR codes absolutely um you have to maintain them you have to clean them yeah you need to have a process of getting them um of having them in a good um yeah clean way otherwise um you your robot might fail might fail to read it okay yeah good that's quite scary so maybe we have to move on and look at some of the more advanced laser uh navigation Technologies because this doesn't sound too promising to be introduced in my location yeah well you are right but think about rich truck operations with heavy loads where you need a reliable navigation and we will come to that uh now if you're just um running around based on Contour navigation or laser navigation you might yeah face other troubles and you don't want that reach truck with a uh 1.5 ton load uh puts the pellet into a wrong position uh uh on 11 M height yes okay I understand there are some reasons why this type of navigation is still used but you're absolutely right it's not state-ofthe-art okay so Contour navigation you already mentioned it which is also known as 2D slam right yes right and this kind of nav ation technology promises us that you don't need to change anything in your infrastructure you just put the mobile robot there and um it will detect based on your unique Contours in your Warehouse shop Flor area where it is MHM maybe let's uh just quickly also talk a bit about this uh uh Hardware that is used mhm so I understand that it's a lar laser basically which looks a little bit like a funnel right if you see it uh in some vehicle and this uh liar is basically taken a slice of the environment uh which is then called the Contour and is using this one slice to navigate but also to locate itself right yes so the ligher means um light detection and ranging so you have these laser beams um coming out of this uh funnel yeah yeah and um you'll get the information back of um what is the distance to okay your to yeah uh to your environment whatever there might be yeah it's like measuring the distance to a star using the speed of light right because then it comes back at some point in time and you can measure okay absolutely okay understood so we have the 2D slam um which dates back to the 2000s I found out right yes and you have this one also probably at home in your um vacuum cleaner robot oh okay yeah it's basically 2D slam what they are using think interesting so already quite common technology that is basically a commodity right yes okay so this is 2D slam now let's move on to the latest technologies that we have and my next one would be fre D slam so what's changed now yeah you I mean 2D slam as the word already implies you will get you you will see your world your environment on a 2d map like looking from the bird's eye perspective yeah that's already good but um imagine you have obstacles on not on the same level where the TWD slam is located usually they are 10 cm High 15 maybe but what if a pet is um standing out yeah on 50 cm 1 M height the device would crash would run into it yeah yeah exactly so with the 3D slam you finally get uh this point cloud of your whole environment in a 3D picture right which is a great advantage since you can navigate and locate yourself with even better precision and safety I would say yes and here um what is interesting and that's why I said the 2D slam is only promising us that you don't need to change your infrastructure because think about um your shop floor on 10 cm height M how does it look like yeah it's a strange position to imagine but the issue is sometimes your shop floor is not unique enough for the mobile robot yeah to know where it is because it might be it looks all the same and sometimes and usually this happens during the implementation and not in the Pres scoping unfortunately you figure out that your shop floor is not has not this uniqueness required and what happens then you need to add unique points so reflector again reflectors it's not a big deal yes reflector tape will be enough um so that yeah the robot can find your environment and understand your environment yeah that's interesting so that means with uh 3D slam um you don't have this problem anymore because it's more the holistic 3D View and I think they also tend to check fixed kind of like um building parts to to be integrated into their uh navigation localization so I don't know like parts of the of walls and parts of I don't know yeah sure you you will have much more information this is part of the story the other one is um the interplay with software so you doesn't matter if it's 2D or 3D if you have a smart software um it might make the difference in localization and um mapping or navigating um because what you will get from the 2D laser is always the data is the same but the question is what do you do with this data and you can use this in a smart way or not so smart way means more or less magnetic uh not sorry magnetic uh reflector tape um um yeah more or less reflector tape at the end yeah good we have one more which is the vlam no that's the newest one that's the latest one uh stands for visual slam and is basically a camera that is able to identify or out information from the pictures that this camera is taking during a drive or yeah during whatever operation which of course now opens up the option for totally new use cases that an mobile robot can take over because it can now for example detect that there is a pallet standing somewhere M that he or she I don't know if it's a he or she but which it could transport or should transport yes this requires uh image processing capabilities so you need a proper processor for that a proper camera so this is also the most expensive solution and you need to think about if you really need it and you need what every camera needs and it's light so in a dark Warehouse uh it will struggle to find uh the p of course but yes this is the the newest way of navigating and it offers additional possibilities as you said because if I have a picture you could think about detecting labels barcodes text and also work with this information yeah um so very interesting technology yeah but maybe to also take again the look at all of these Technologies um what we also see in the market is that there is a lot of mixing of different Technologies in one vehicle or in one mobile robot because um they have different strands right so for the slam navigation it's a must if you want to have obstacle avoidance so if you want to really operate an AMR you won't be able to do it with Magnetic Tape but the drawback is the mobile robot is or might do crazy things and the Precision Precision is not as high so it's roughly around 2 cm plus minus okay which is a problem when you have a station a source and that you need to service with the mobile robot this 2 cm plus minus which you have with slam is a big problem right because he might not hit exactly this source of destination yes especially if it's a under rer for example which needs to really go beneath something to to lift it um and this is why we often see a combination of Slam but then still some line guidance or Magnetic Tape which is then actually used just in front of the source or destination for the mobile robot simply to en able a very accurate um dock in and dock out process at this particular location yeah because these are the typical situations of mobile robots where they lose a lot of time if they are struggling to find the exact position to pick up a pallet and yeah you can avoid this by having at the source and destination a magnetic tape or any visual tape yeah yes at the end from my point of view the most successful agv AMR vendors will be the ones understanding best the interplay between the data they are getting from the sensors and the software they are implementing that's a very interesting statement but I can totally go with it um I think it's not going to be this one particular navigation technology which will be a one size fits all it's rather going to be a mix of different um navigation Technologies localization Technologies which need to be brought together in a Smart Way um to get the best result and you know get the job or the requirement fulfilled yes and as always it depends on your use case do you really need all of this Tech technology because you have to pay for it every single sensor camera which is attached to the device you have to pay and it's quite expensive right when we look towards the slam technology that's not peanuts that's really proper money um they cost and if you compare a line guided vehicle with a slam guided vehicle and some providers do it even right yes and there is a quite big difference actually yeah the price could be even doubled yeah yeah so know your use case know what you really need and um don't over engineer it um just really go with the navigation or localization technology which will get your job done and um if you struggle to find the right solution again we suggest to talk to some experts some Consultants to really prevent you from Buy stuff which is great but which is not really a must for you and it's just costing more money absolutely a great conclusion of this episode Matias I feel like I'm the one to do the conclusions the whole time no it's uh uh it's basically it's basically it and um let's shout it out and um make sure that uh all of you guys are well informed so was a pleasure to have you on this week's episode um hope you learned something and got some new insights and uh we of course hope that you're going to join us next Monday again for another episode of the automation Awakenings podcast um until then all the best take care and hear you soon hear you bye-bye this was another episode of the automation Awakenings podcast visit us at automation.com [Music]
2024-05-05 10:01
