Latest / Autonomy Insiders / Baidu Apollo Robotaxis in Switzerland: Inside PostAuto's AmiGo Project

Transcript

Baidu Apollo Robotaxis in Switzerland: Inside PostAuto's AmiGo Project · Autonomy Insiders

• Daniel Abreu Marques: Welcome to Autonomy Insiders. Today I'm joined by Martin Neubauer, head of Autonomous Mobility at PostAuto AG Martin has been working in autonomous mobility for many years. He was also managing director of the Swiss Association for Autonomous Mobility and he's also a board member at PAVE Europe. So he brings a long track record in autonomy across industry, policy and operations. And today we want to take a deeper look at Welcome to Autonomy Insiders, the show where global industry leaders in autonomous driving unpack their real-world insights. Post Auto's driving strategy, what they're building with their AmiGo project, and how autonomous mobility can realistically scale in Switzerland and Europe. So Martin, really great to have you on the show. Today with Martin Neubauer. Martin: Likewise, hey, thank you so much for inviting me. really a pleasure ⁓ be able to join today and answer your questions and get in with you as autonomous mobility part of my history, let's say for the last up to 10 years actually. So for quite some time now. And love to talk about it as you can imagine. Daniel Abreu Marques: Yeah, absolutely. Me too. let's start. I think some of our listeners may not be familiar with PostAuto. And could you briefly explain what PostAuto is today ⁓ and also the it plays in Switzerland's mobility system? Martin: So ⁓ PostAuto is the biggest or largest public transport operator in Switzerland, especially in rural, ⁓ alpine and peripheral regions. So we are basically the backbone of a public transport to serve certain accessibility, especially outside of major cities. We are part of the Swiss Post, one of the biggest logistics companies here in Switzerland. We are fully integrated of course into Switzerland's public transport system national timelines, unified ticketing, guaranteed As you know, Switzerland is quite well known for having one of the best public transport systems worldwide and we're to be a part of it. ⁓ Daniel Abreu Marques: why autonomous matter to PostAuto in the first place? So what problems are you actually trying to solve by engaging with autonomy? Martin: So first of all, autonomous mobility or as we call it here in Switzerland, automated mobility not basically a new topic for us. We have ⁓ invested into autonomous mobility or automated mobility for quite some time now. We started in 2016. ⁓ We actually the first ⁓ public operator worldwide using autonomous vehicles at that point from the French manufacturer ⁓ Navia. and use them public roads. We had lots of visitors worldwide, Google for example, Toyota and you name who had a look at ⁓ our different projects that regard. And from that moment on, we started to invest further into autonomous mobility. think, ⁓ or we think one of the main reasons why we have to go this path ⁓ is it's different or different reasons. a shortage of professional drivers at some point. depends with kind of PTO you're talking to in Switzerland or in Europe. But more or less, we all have the same issues at some point. ⁓ They are retiring a certain percentage of them, right? It has become complicated to find new bus drivers, ⁓ especially professional ones regards to rural This is clearly one of main reasons why we ⁓ invest into autonomous mobility. see rising costs ⁓ combined with a certain political pressure also maintain affordability, putting aside high safety and the high service quality level that we have. And being a rural public transport operator, we also face economically unsustainable or unsustainability in low demand regions where we are operating. ⁓ autonomous mobility, at least currently in combination with an on-demand setup, is for us the right way to go. Not saying that we have a business case. Any time shortly, at least, But nevertheless, we also see that from a market perspective, if you look at players such as Uber, guess most of your listeners have read that Uber yesterday that they are ⁓ setting services in different cities, also including Zurich. Let's see when this is really happening, ⁓ perhaps 2026, perhaps a little bit later, I don't know. But these kind of players are also entering the market, right? And we also need to gain knowledge and to build up knowledge and to build up know-how in that region and that area. So these are basically the main drivers for us to invest into automated mobility. Daniel Abreu Marques: ⁓ And do you that PostAuto as a public transport operator has a ⁓ special responsibility when it to experimenting with these autonomous mobility solutions? Martin: I think due to the fact that, of all, also being part of a company, of course, we have certain freedom to ⁓ do testing that ⁓ area. Smaller I for them, it's more complicated to money into ⁓ that area. One of main reasons why we have founded the Swiss Association for Autonomous Mobility, right? Creating ⁓ not just a platform, an ecosystem. sharing knowledge also getting an understanding of what it means for our mobility system, not just in regards to a PTO or public transport, but also in regards to companies or privately owned companies who have little bit of different interests. But clearly for us operating in rural areas and in these regions, ⁓ ⁓ Knowing that at some point it'll be complicated to keep up the high service level, need to do something, right? ⁓ that's one of the main reasons ⁓ I why we are invested into that region. Daniel Abreu Marques: So you mentioned that PostAuto is also a logistic business. So is your autonomous strategy also going into the logistic business with autonomous solution, or your primary focus is just on the public transport side? Martin: For now, it's only public transport. We had trials ⁓ logistic vehicles, but that was ⁓ years years ago. So for its focus is definitely public transport only. But we are in exchange, of course, to understand how to approve such on public roads. But for now, it's only public transport. Daniel Abreu Marques: All right. Then let's talk about recently first major announcement, I would say, in autonomous with the AmiGo project and together with Baidu Apollo. Can you walk through what this project is really about? So what is exactly being tested, built, and learned in this phase? let's ⁓ talk about AmiGo project with Baidu Apollo. Martin: So main reason or how we actually started is that we had regular on-demand services with regular vehicles without autonomous vehicles. And we see ⁓ a lot of potential in setting up these kind of services, knowing of course they are expensive also for basically the cantons who we have to these services. So main idea why we want to go in that direction is to build up these kind of on demand services in so-called peri-urban areas to rural areas. Especially we are currently operating in Eastern Switzerland with our AmiGo vehicles or start to operate soon, so close to the border of Germany and Austria. Main idea, we see it in different phases, is at least for now is validating the technical feasibility. Swiss conditions, ⁓ fog, we have rain, ⁓ snow, obviously as well. We want to see the demand on the customer side, the acceptance also on customer side, and see how we can further push basically the topic at some point in other regions, of course. Knowing that we have operations with asset in the beginning, other vendors, But it's something different if you operate or start operating with three vehicles. Now it's 25 vehicles. At some point, it'll be more, right? And this understanding costs, of course, time. need to build up that kind of knowledge. It's completely something else if you operate three vehicles, if you go up to ⁓ 25 or even up to 50 vehicles. And from our point, as already said, we have to invest time and money. in getting this kind of understanding how to operate these vehicle or amount of vehicles. And that is also clearly part of phase one basically for us. Daniel Abreu Marques: And you already mentioned that currently it's capped at 20 or 25 vehicles. ⁓ But you want to scale if there is enough demand and ⁓ if everything works well. ⁓ do you have an number in mind how big such a fleet could get ⁓ for kind of service in Switzerland? Martin: You Well, that's a hard question to answer. So there's this rule of thumb basically that per 1,000 inhabitants, you have actually one vehicle that really rough, like really rough to be honest. Of course, we have been doing lots and lots of simulations. But in the end, it's something ⁓ you simulate a lot, but real life always looks different, let's be honest. So it's... No matter what kind of number I will tell here and to you and your listeners, it won't be the right number. We will see, especially getting an understanding of how to operate, how many shifts are necessary, how the demand, for example, is at night. We will then see how many vehicles are actually necessary to cover a region that is 50, 80, 150, 200 square kilometers big. Daniel Abreu Marques: Right. made Baidu Apollo your ideal partner for PostAuto? Could you walk us through how collaboration came about? And also maybe how you're bridging gap between a Swiss public transport operator and a Chinese autonomous technology provider to kind of ensure a smooth project rollout. Martin: Well, let's be honest, the second question was a big question for us to answer internally, of course, being part of the government, actually, as PostAuto is and collaborating with a Chinese manufacturer. We also expected a little bit more pushback from the media, which didn't happen at all. There were questions, obviously, but there's a lot of trust towards PostAuto ⁓ as we are a company with a long lasting history also in regards to trust. I think this was so much of an issue in the end. One of the reasons to answer your first question is we of check the whole market worldwide to get an understanding who's ⁓ a capable partner for ⁓ to such a service. ⁓ they are not... Currently, at least not a lot of them, right, who look into the European market, who think the European market is suitable. Also understanding the different legal or the regulation that you have, let it be Germany, Switzerland, France, Austria, and so on and so forth. And clearly one of the reasons for Baidu was that they have an extensive real-world experience with, let's look at Wuhan, for example, with a large-scale operation. So they know how to... First of all, scale, they know how to operate in 24 seven, basically. of course, also have been to ⁓ to talk to them, to check the operation ongoing, ⁓ look a little bit behind the curtains. And for us, this was clearly one of the main reasons seeing that there's a ⁓ L4 technology running on a daily basis, ⁓ which is not just the tech itself being proven but also accepted by basically the end customer and see how the end customer ⁓ the service works. And it interesting to see that ⁓ it's special anymore on their public roads, most probably also the same of course ⁓ for US. But these clearly one of the two or three main reasons why to go ahead Baidu and also the way they behaved. and the way they forwarded information. said we took a lot of or invested a lot of time into finding the appropriate partner. And was top notch for us. I forwarding the information, being transparent and open minded, also understanding ⁓ how is behaving, course. That ⁓ was in the pretty for us with whom to go for. Daniel Abreu Marques: And Apollo is also bringing its latest generation level 4 vehicle for AmiGo Switzerland. Are there any or specific testing that is required now to ensure this vehicle also handle the Swiss roads and conditions? You mentioned earlier, it's snowy, it's foggy, so Baidu isn't used to it right now. Martin: Correct, of course we have to adapt. China is not Switzerland. I'm not saying that their road conditions aren't complex. They have a completely different mixed traffic than we have in Switzerland. I would say people in Switzerland behave better in the traffic and obey the laws that we have in traffic in Switzerland. So this is making our life a little bit easier. But as I said, we are a rural public transport operator, also working in peripheral regions. So you do not have a lot of reference points for your lighters, example, cameras, and so on and so forth. So clearly something we have to test and see how the vehicles are behaving in these kind of areas and scenarios. Nevertheless, also, the idea of the project is to get a certain technical understanding. How do they behave in snow? ⁓ as you can imagine, the main idea is not to operate on public level from day one on. We will do a testing of the vehicles in different conditions together with the federal office. We are still currently operating in a so-called pilot framework. We have an ordinance or a new kind of regulations with Switzerland that actually allows us to ⁓ a regular service. But that still something ⁓ where the basically have to approve it. In our case, it's the federal office of roads that will give us a ⁓ so-called exemption permit that we can at some point operate without safety drivers. So there's a certain ⁓ behind it that we have to follow through together with the federal office to that at some point then that we get basically the safety driver out of the vehicle ⁓ and see we have a smooth and safe operation. But clearly we want to test under rain, rainy conditions, foggy conditions. I see what it means if you have ice on the roads, how the vehicles behave. We are also working with the TCS, the Touring Club Switzerland. They also have test tracks ⁓ where can operate in ⁓ a under conditions with these vehicles. So it's all about testing, testing, testing to get an understanding of safety. Safety is our highest, ⁓ or most important point for us, as this is also the expectation by the end customer, that we have to be safe, but also of course for us internally. Daniel Abreu Marques: Absolutely. let's stay a bit at the safety driver topic. Do you already have an indicative timeline for when it could be possible to remove it? Are there specific conditions ⁓ that have be met before PostAuto would feel comfortable taking that steps? Are there milestones or something like that? Martin: So, said, safety drivers are clearly of importance as we are still in the testing phase regarding let it be a risk management regulatory compliance due to federal and also for us to understand public acceptance, but we do not yet have a fixed timeline saying, okay, on that date, we will get rid of the safety drivers. I said, if we have the feeling together with the federal office, but also with Baidu and our internal management, Now is the right timing to get basically the safety driver out. Let it be by beginning of 2027, Q1 2027. So we want to use the full year for testing actually. And as I said, it also depends on the permit from federal. We have certain stages that we have to fulfill showcasing a high level of safety. And that is something we are going through at some point. And that will also take some time. So let's see. probably beginning of 2027, but perhaps it can also be a bit later. Daniel Abreu Marques: Sounds good. And one topic that is often also discussed when we talk about autonomous vehicles is ownership. In the AmiGo setup, who actually owns these autonomous vehicles? Is it PostAuto owning and operating the vehicles, or does Baidu Apollo Go retain ownership while PostAuto focus is more like on the operational side? Martin: No, we own everything. We own the vehicles, we own and set up the operation, and we also own the data. course, according to ⁓ data protection laws, let it be for personal data, but also for mapping data, for example. One of putting aside that China, doesn't actually matter if it's China or ⁓ the vendor is from another company, as it we own everything. We own the vehicles, we own ⁓ set up the operation, and we also own the data. being part of a governmental company, also have certain internal guidelines that we need to fulfill. And this was clearly one that we have data ownership and understand also the data. Daniel Abreu Marques: Baidu also often highlights ⁓ how cost their robotaxis are in China. And you're looking operating these vehicles then in Switzerland, do you think these cost structures are realistically be transferred one to one to your operations? Or do you see meaningful adjustments and cost uplifts because it's the European context and not China anymore? Martin: So I wouldn't say that a direct one-to-one transfer regarding the China costs is doable for us in Europe or in Switzerland. We have other cost drivers. We have completely different labor costs in Switzerland, as you can imagine. We have completely other regulatory requirements that will increase the costs. We have... Perhaps other standards regarding safety and redundancy standards or understandings. said it doesn't matter if they're from China, US or somewhere else. We have own standards structure in Switzerland that makes it hard to ⁓ compare cost drivers. it might be bit too early to also talk about ⁓ costs as it for us currently focuses ⁓ on learning effects. They important to us to understand the costs also on our side, ⁓ what means. think ⁓ ⁓ is how to see how the regulatory requirements will adapt and change. Yes, we have an ordinance in Switzerland which has come into force ⁓ by March or last year in March. And let's see this will be also adapted throughout ⁓ the time and and how it will adapt also the costs in our case. For example, the ordinance says the The operator, as we call it here in Switzerland, so that the person basically surveilling the fleet the back office, that person has to sit in Switzerland. at Some of them are on the Philippines. The operator, as we call it here in Switzerland, so that the person basically surveilling the fleet the back office, that person has to sit in Switzerland. I most of ⁓ a of your listeners also heard it. I think was yesterday or two days ago. The operators, I know how they call it, how they call them ⁓ at Waymo. Some them are sitting on the Philippines. Not understanding, not fully understanding the setup from Waymo, of course. ⁓ But that is something we are not able to do here in Switzerland and also don't plan to do. But that is something we are not able to do here in Switzerland and also don't plan to do. So we have different cost drivers. Daniel Abreu Marques: OK, so you would basically then also have your own remote operation center where you can then kind of like support the vehicles. So it's all ⁓ Swiss based. Martin: Yes. Of course. Of course. Of course. Daniel Abreu Marques: And you also talked about ownership. And how are you then, as PostAuto, approaching the data governance and the cyber safety in the project? So how can you make sure that data stays in Switzerland? Martin: a ⁓ advantage of being part of a company having certain departments for cyber security, but also data security. That is their business, right? ⁓ To confirm exactly kinds of questions. We of course also obey to Swiss data protection But as you can imagine, working with such a vendor, it's important for us to also understand the data flow, right? So we created a certain ⁓ infrastructure that will help us to also ⁓ that stays in ⁓ or in Europe. yeah, but still something ⁓ also ongoing that will see also to the new ordinance, what is expected in that region or... in that aspect. it's an ongoing learning process for us as well, but also for Baidu of course. Daniel Abreu Marques: Makes sense. Martin: But including, I think that is also of importance. not just you buy these vehicles, you put them on the road and ⁓ is happy. Phase one for us is clearly about learning. So we will do ⁓ tests. We will see how the, example, if you ⁓ have ⁓ updates, there's a certain process behind it, how to understand what the software update is about. how to improve it also from our side. that, guess, also will be part of the homologation and type approval process at some point. But as I for us, that's currently not the case, as these vehicles don't have type approval. ⁓ And there no type approval currently really existing in Europe yet. But we see that we have certain measures to ⁓ be exactly that site. Daniel Abreu Marques: And could you us through how this actual approval process for the AmiGO project went? So from the idea to getting permission to operate, what did the journey look like? Martin: The journey itself is not new to us and it hasn't really changed actually or has been adopted, the idea said it started cooperating with the federal offices in 2016, 2017. We have to prepare different concepts, safety concepts, communication concepts, route concepts. We also need to showcase data in cybersecurity. Training courses, for example, is also something of course of high relevance. How do we train our personnel? all these kind of different concepts that we and assessments that we have to fulfill in cooperation with the federal offices and also of course with Baidu. ⁓ And ⁓ the process itself, we need to showcase that there's a high level of safety. So we will have certain test scenarios that we need to fulfill and showcase towards ⁓ federal office that we basically check mark these different test scenarios. And yeah, you imagine that takes some time to... the documents but also to go through this approval process it's not just handing in the documents and then you are allowed to drive without a safety driver. No, there's a lot of testing behind it on areas, ⁓ of on public areas. also include the municipalities, we also include the cantons, we include the police, ⁓ the So it's ⁓ actually creating ecosystem. ⁓ involve all the necessary players. So not just training of our personnel, it's also training of the personnel from the police ⁓ ambulances really a high safety standard. And that is where the federal pushes us, which is good. But ⁓ as I we have internally high guidelines on to be And is basically how the looks like. ⁓ high level, a high level. ⁓ Daniel Abreu Marques: Right in the process doesn't change ⁓ you have like a Chinese ⁓ driver technology partner. So there are not more documentation requirements or no adjustments to the process compared working with a European partner. Martin: I wouldn't say so. think there's a... Most probably, of course, due to the fact that you have other regulatory aspects in China than in Europe. I don't know the setup, how it works in China, of course, and what they need to showcase that they the permit. But I don't many differences. I wouldn't say that there are a lot of differences. Never really worked, of course, with a European player, or not yet. ⁓ I think it's more like cultural barrier that need to adjust to, right? It's a different culture, of course, but not a really regulatory ⁓ side, just it's Chinese player, we have to hand in more documentation. ⁓ is not the fact. Federal is pretty sticked on receiving certain documentation. ⁓ But together with Baidu, we're handing in this kind of documentation. not really. I would say there's a difference to any other player. Or federal makes a difference to any other kind of player just because it's China or not China. For them, it's about safety. doesn't matter where the player is coming from. Daniel Abreu Marques: Understood. if we imagine everything works really well and you need to scale the service, will this be an exclusive kind of partnership with Baidu or will there be more partners ⁓ involved time if everything works well? Martin: Well, for now, the partnership is about up a service with 25 vehicles. And see going to happen afterwards. I said that the collaboration currently is really well ongoing. the service itself works well, then course we can imagine to scale further with Baidu. ⁓ But it's hard to see how the market is behaving, And a lot of change lately, lots of different corporations, especially you look at companies such as Uber, who's currently, I think, working with 20 other current ⁓ players We, of also as Baidu itself, open-minded to also at some point with other vendors or stakeholders. But for we tried to and set a service together with Baidu with 25 vehicles. then we have three years of time now ⁓ then see how we can perhaps ⁓ collaborate together. Daniel Abreu Marques: And you mentioned that the service is intended to complement existing transit network in regions where we have sparse service. And how do you see then the autonomous shuttle integrating with the traditional, for example, bus routes or the general public transport system overall? Will this more like be a first and last mile connector feeding to the main lines or even replace, for example, ⁓ low ⁓ bus for example? Martin: The advantage is that in that region ⁓ don't have bus lanes ⁓ we need to replace. So it's basically offering a service in an area where public transport is not really existing or existing to a point that it's good enough to set up such a service. But in regard, we also did lots of simulation also in other regions, of course, and thought through if. placement is making sense, is not always the fact you was just using autonomous vehicles in a on-demand setup. But that it's clearly a last first mile feeder basically to certain points of interests that it be at next train station, for example. So being a public ⁓ transport operator, of course, of high value for us that it will become an mobility something we also try to constant work on. Daniel Abreu Marques: I also noticed that ⁓ PostAuto plans to hold information events for residents before launching the tests. And what the keys to engaging the local community and maybe turning also initial skepticism into Martin: So we have done a lot of them and of these events local residents. So course we want them to use the service and to give us feedback regarding the service. ⁓ So we this kind of communication layer quite early, especially the vehicles appear on public roads. We are a public operator. We have a certain understanding and in the market. But for us, it's also to start open dialogues. What are people fearing? What do they expect from us also ⁓ the near time future? And until now, the feedback was really positive. Of course, we discussed certain topics such as ⁓ ⁓ impact jobs, for example, transparency, when the pilot will start. all of these different aspects. But interestingly speaking, more than, I guess, 90 % of the people were looking forward to the project, asked us, when can we finally download the app to use your service? When are you going live and public? So the feedback until now was really positive. And we stay in close connection with the different residents. as we have to implement them, of course, and gain further feedback and improve the service together with them. That's quite of importance for us. And also understand when something is not happening the way we want it to be. So here this ⁓ connects with the residences of importance for us. Daniel Abreu Marques: And ⁓ talked ⁓ the app. Is this then a special app only for the autonomous service or is it kind of an integrated app where you have also like bus and tram and everything inside? Martin: now, it's ⁓ ⁓ special app, or basically only used for the on-demand service itself. And throughout the whole process, we will then see how to be able to ⁓ integrate it into the regular system. Daniel Abreu Marques: Then maybe let's look beyond Switzerland. do you think Europe needs to get right if autonomous mobility is to move from pilots to ⁓ deployment? Martin: It's an interesting question most people will say ⁓ Europe is behind. Yes, of course you see what's happening in China, you see what's happening in ⁓ US, you see that all of these players are scaling a but ⁓ they have different systems, For sure, we need autonomous mobility and we discussed different points why autonomous mobility in Europe is also of importance. But I'm actually more optimistic than cautious to be honest, knowing that of course autonomous mobility has to be designed for the European market. I think it's important to say that we... look heavily, not just us in Switzerland, all the players in Europe, in Germany, in France, in other countries such as Norway, ⁓ it comes to safety, sustainability, and course ⁓ inclusion. We are taking more like a deliberate, ⁓ perhaps different path, of course, than other countries. For us, The goal is not just to deploy as fast as possible and scale to prove there is a business case, but to deploy as safe as possible, basically. It's also a continuation of ⁓ this vision zero kind of idea that you would like to reduce also traffic accidents or human errors. That's more the kind of leadership I also expect from Europe in that regard, not just scaling as hard as possible and as fast as possible, but also as safe as possible. So it's not a winning at a race, but it's more doing it in the way that it's safe, integrated, but also, of course, being able to scale it. So we are also from the discussions I have with lots of other PTOs or other. other mobility providers. all want to create a certain safety culture. We want to integrate it into the that we have. And yes, understood that speed and scaling is an issue in Europe. But I wouldn't say that we are in that regard slower, right? just that ⁓ aspects are making us, and financial aspects, of course, obvious as well, ⁓ making us slower. yeah, will come to that point, I think. you see it from what's happening in Hamburg, you see it of course from what's happening in Switzerland, Paris has or as France has also becoming really active, other cities in Germany as well. So there's a lot happening in the market and ⁓ pretty proud of that to be honest. Daniel Abreu Marques: Yeah, absolutely. ⁓ we also notice ⁓ and more interest from ⁓ the driver companies to come to Europe. So do you have an opinion why that is the case right now? Martin: You mean for now that these companies are now looking into the European market, the European market itself, of course, is big enough it has to be conquered, of course. You see that ⁓ also a regulatory side, that putting Switzerland, also perhaps looking into Germany, that we have legal frameworks in place. Daniel Abreu Marques: Exactly. Martin: something that is not happening for other countries or not yet happening for other countries, which has to happen at some point. I think and also how the is behaving lately and that people also on a political level, you also need political pressure, of course, that there's also a lot on the on political level, EU political level. that people see Europe as an market basically that is ⁓ ⁓ ⁓ Daniel Abreu Marques: if you change one thing at the European level tomorrow to accelerate autonomous mobility, what would it Martin: I think reduction national Yes, we have certain legal frameworks, but they have to be adapted, of course, which would give us as an operator ⁓ room to ⁓ able to ⁓ easier for the also not for us as an operator, but also for the suppliers or the tech providers. a clearer political commitment beyond experimentation. We have so many projects or so many pilots. Please let's now to the step, right? It's not, it'll not help you at if you have another project with five vehicles or less or a bit more really understand the processes needed in the background to set up such a service. That definitely has to happen. Cross-border would be insanely interesting to see how that actually works. That something I would love to see at some point. I know some countries are in discussions in that regard. There are also some potential projects upcoming. And in the type approval, let's be honest. And ⁓ actually harmonized approval process for autonomous vehicles. So it's more or less these four topics or four points I would see that would help the market a lot here in Europe. Daniel Abreu Marques: Absolutely. I agree. So we from, for example, Waymo, they have these large depots. And you already have those depots, or you need to build a real depot for these 25 or then in some time even more vehicles. Martin: Of course, as a PTO, we have depots, but we are not a city public transport operator where you can easily store like hundreds of vehicles. So normally in our case, we store 10, 20, 30 vehicles max. Coming to your question, we had to rent a depot basically and adapt to our needs. But also, hindsight, see that, of course, we can scale to more vehicles that was of importance for us. we the ideal depot, basically, for us also ⁓ to able to scale. So we had to come up with something new, actually. ⁓ Daniel Abreu Marques: perfect and And the service is then available 24-7? What are the operation times? Martin: ⁓ Regarding testing, of course, we will see how the vehicles behave at night. It's more or less a demand question. If we see there's a certain demand, as you can imagine, of course, it ⁓ comes a certain cost, right, ⁓ to the person working at night as well. We will see regarding the demand, if there is a certain demand, then we might go to adding another shift. But currently, we start early in the morning. We will see that we can cover the higher frequencies to ⁓ some point night. That it be 10, 11 o'clock. But again, ⁓ really depending the ⁓ demand that try to establish. let's see, perhaps we have something also at night. Daniel Abreu Marques: Do you have rough ⁓ ratio that you have from a remote operator to vehicle? So it one to one, or one to three, one to five? Martin: we will, as we don't start with 25 vehicles from day one on for for example, we have a one to one ratio, knowing that will not really help us regarding the business case, of course, we will see how to further adapt also then at some point receiving more vehicles from Baidu how we can adapt to increasing or ⁓ adapting basically the from one to something. But that goes ahead also with the regulatory aspects or the expectations from federal to us how fast we can improve the system together with Baidu. Daniel Abreu Marques: are also sometimes discussion the levels of ⁓ remote operation. are you to drive the vehicle? Are vehicles just requesting, OK, I'm now here in this scenario. I want to do A, should I do B? And then you choose A or B, and then the vehicles themselves? Or is it more like really taking over the vehicle and driving it? Martin: we are not allowed to influence the dynamic driving test of the vehicle. So on a strategic level, for example, perhaps to PUDO, that is basically something we are allowed to, but ⁓ really ⁓ the driving task. So let's say steering it or hardbraking ⁓ the vehicle ⁓ something That is for us not Level 4 So not something that we will push or pursue. As I said, on a strategic level, it's something else where don't intervene basically in the DDT. But that's us real level four we would like go to. In that regard, of course, for this, we onboard But that's about it. We not intervene in the Daniel Abreu Marques: So this was a very, very insightful conversation with you. And I you really, really much for being here and sharing all this experience with us. Martin: Thank you so much for having me today and being able to provide certain information on our AmiGo project here in Switzerland. think Switzerland is making a lot of progress lately. For us on different levels, autonomous is becoming mandatory, knowing that it will take a few years, right? But there are also, as you see yourself, other players coming to the market. yeah, they're really looking forward to the near-term future and see also in other countries how Autonomous Mobility will further develop. thanks again for having me. I really appreciate it and cannot wait for my panel talk and interview with you. Daniel Abreu Marques: Perfect. We're all looking forward to have you on back again. Thank you very much, Martin. Martin: Thank you so much. Thank you.