2970: How LeddarTech is Leading the Way in ADAS and Autonomous Driving

[00:00:01] What does the future hold for advanced driver assistance systems and the world of autonomous driving technologies? Well today my guest is the president and co-founder of LeddarTech. They are a company at the forefront of the AI-based

[00:00:17] sensor fusion and perception software. So today I want to learn more about LeddarTech's innovative technology, how it's transforming the automotive industry by enhancing the capabilities of ADAS systems and how it's making autonomous driving safer and more efficient. But what does ADAS stand for?

[00:00:37] How is technology influencing the decisions that automakers are making today and what can we expect as it evolves? These are just a few of the questions I'm going to answer today. We're currently producing something like 30 to 35 episodes every single month reaching around about 130 to 140

[00:00:56] thousand listeners around the world. So I wanted to mention the sponsor of Tech Talks Daily and also shine a light on how legacy managed file transfer tools are dated and lack the security that today's remote workforce demands. And

[00:01:10] companies that continue relying on that outdated technology put their sensitive data at risk. But the good news is you can ensure top-tier security with Kiteworks MFT suite. Their software security hardening includes an ongoing bounty program, regular penetration testing to minimize vulnerabilities, the

[00:01:29] convenience of a one-click appliance update and avoid the hassle of managing multiple unhardened servers. And remember Kiteworks is also FedRAMP moderate authorized. So step into the future of secure managed file transfer with Kiteworks by visiting kiteworks.com to get started. That's kiteworks.com

[00:01:49] to get started today. But enough rambling from me let's get today's guest on. So buckle up and hold on tight as I beam your ears all the way to Montreal where my guest is waiting to join me. So a massive warm welcome to the show. Can you

[00:02:06] tell everyone listening a little about who you are and what you do? Well Neil pleasure to meet you. My name is Fran St-Hilaime. I'm president CEO of Lettertech and you know my background is all been in the high-tech industry. You

[00:02:24] know I've got 25 years of experience in the industry really starting from semiconductors all the way to sensors to system software to now AI. So I've gone full spectrum from your in-car infotainment. This was back then your

[00:02:42] telematics platform in the early 2000s late late 90s early 2000s to the advent of audio where you had 16 speakers in your car, 22 speakers you know your Becker, your Levinson's, your Lexicons you know these high-end systems Bose and so

[00:03:01] forth. So I went grew up in the industry developing semiconductors and software for telematics and in-car infotainment and then progressively migrated to more safety developing sensors for safety applications. So so things like your air bags they need sensors, things like your rollover tilt sensors you know they need

[00:03:27] sensors as well but also braking brake pads they need pressure sensors you also need median NOX or gas sensing because you don't want to suffocate inside the cabin so the air has to be regulated based on sensing as well your HVAC

[00:03:44] system sensing. So that's my journey and now today I'm at Lettertech. You know it's really kind of the end or the latest foray for me in the automotive space. Well it's a pleasure to have you on the podcast because every

[00:04:02] single day I try and get people thinking differently about technology how it's impacting their life, their world and often in areas that you don't automatically associate with technology and of course you mentioned cars there

[00:04:15] and all those different sensors and cars are for the most part like almost huge computers arguably even data centers in some situations. So I suppose just to set the scene for our conversation today and bring to life what we're talking about

[00:04:30] can you tell me a little bit more about Lettertech's sensor fusion and perception software and how that's helping improve the safety and accuracy of advanced driver assistance systems and autonomous driving. It's an incredibly cool space

[00:04:45] that I will hear about it but a little bit unsure I'd love for you to demystify and tell us a little bit more about that. Absolutely but first a little bit about Lettertech. Lettertech was founded in 2007 as a it was a product of

[00:04:59] the university research so founded in you know seven as a sensing company developing these unique solutions from signal acquisition so acquiring the signals from from the environment to processing those signals into the digital domain and we use them to make something called LIDAR so you have

[00:05:23] radars which are radar radio detection types of signals so radio signals that are used in detection and ranging and then you have LIDARs these are light light beams or light signals that are used for detection and ranging and so we

[00:05:42] were one of the pioneers in that space. Our products were used for speed enforcement and you know toll bridges and things like that but also in agriculture or construction or mining and it was a pretty successful business

[00:05:59] but but it was a very fragmented market and so as a result you know back in in 2015 and and progressively into 16 and 17 the company evolved to become a pioneer in the field of AI based fusion and perception software so what is fusion

[00:06:22] perception software it's essentially you know the car needs to be aware of its environment in order to be able to aware to be aware of its environment it needs to have a view of that environment so Lettertech's software basically we use

[00:06:43] you know AI and computer vision to fuse all signals coming from the different types of sensors from cameras radars ultrasonic sensors or or even LIDARs as they come into the vehicle and then we combine that with the vehicle dynamics

[00:07:02] so the motion what's happening with the vehicle is it driving on gravel is it or on asphalt is it is it you know in traffic or not or driving at full speed so the vehicle dynamics tells us a lot about its environment and so we fuse all

[00:07:19] of that to recreate the most accurate 3d model of this environment and we do that up to 30 times per second to enable the vehicle to make safer decision and to be to have a more accurate navigation this enables advanced driver assistance so

[00:07:43] things like for example emergency braking adaptive cruise control or eventually to autonomous driving so you need to have this this this call it environmental model in order to have the vehicle think about it as its eyes and

[00:08:03] ears on the road so instead of you as the driver having pretty much a pretty narrow field of view you now have the vehicle with a 360 degree coverage of its environment that's reproducing this environment 30 times every second that

[00:08:20] allows you to detect vulnerable road users like pedestrians or cyclists or or other cars or other vehicles to avoid these collisions or avoid you know unwanted accidents wow it's incredibly cool and when I was doing a little

[00:08:36] research on you before you came on the podcast one of the things that really stood out to me is how you pride yourselves on being sensor and processor agnostic so keep telling a bit more about that and the benefit it will have

[00:08:49] for automotive manufacturers and suppliers who use your software solutions because again it seems like a huge deal from the outside looking in indeed in fact this is quite significant you know for decades the automotive industry assumed that you know accidents were inevitable and so the industry

[00:09:17] evolved to developing you know mitigation accident mitigation technologies like your airbags like your seatbelts for example basically all the efforts were put into you know mitigating the the results or the impact of an accident and so over the past ten years regulators across the globe have

[00:09:46] recognized that change is required in regulations and and so basically regulators pretty much everywhere are working really really hard to basically eradicate accidents and and these regulations you know are somewhat you know in the European Union called the general safety regulation enacted in

[00:10:17] July 2022 and more recently Nitsa in the US called Nitsa 29 which is in a nutshell you know says that cars have to be equipped with enough sophistication enough intelligence to recognize a vulnerable road user regardless of the environmental condition and to have enough intelligence to have emergency

[00:10:44] braking or emergency steering to avoid the collision altogether now this is important because this really essentially go forces the cars to go from safety warning systems to active safety systems and now this is a big

[00:11:03] difference so warning you can have a large margin for error because you know you don't have to know if if what type of object is in the is in the blind spot just need to know there's something in the blind spot and and you don't have to

[00:11:22] know what if it's a car that's in front of you or if it's a pedestrian you know if it's a car you could have a fender you know if fender vendor but if it's a pedestrian you can't and so these vehicles have to have enough intelligence

[00:11:40] current systems current technologies use our camera centric in general and they use what is called object level fusion and these object level fusion is used to analyze the camera data and they're insufficient typically they come from

[00:12:03] the camera provider so whoever's making the camera in this case like a Mobileye they will use the camera and also the radar provider will also use its own radar to provide intelligence so each of these systems independently decide is

[00:12:20] this a pedestrian or is this a car is this a paper or is the debris on the road or a tire debris and it typically will have inconsistencies and most of these vehicles have difficulty resolving these inconsistencies that's why the industry

[00:12:42] is recognized a new approach is necessary we on the other hand later tech we we use what is called AI based fusion and perception where our software combines all the output from all the sensors without filtering without making

[00:13:00] a decision by fusing all of that we recreate a 3d model of the cars and environment and as a result it's more accurate it also allows the car manufacturer to actually mix and match the type of sensors that are needed for

[00:13:19] the different geographies the different vehicle types if you buy a Bentley it's very different price than if you buy a Fiat and so having the ability to have a software that is that is recognizing that does not use complex hardware

[00:13:38] complex sensors or complex processors for a Fiat is very different than if you're a Bentley where you know performance is your clear clear objective and so you need to have better performing at all at all at any not any cost but you know

[00:13:57] cost is less of a problem and so our software stack is a unique software stack that allows basically the manufacturers to do that but also the scalability necessary to have one single software that can do your basic car models to

[00:14:15] your premium car models in fact we have demonstrated that we can deliver twice the performance at almost 50% of the sensor cost using that same stack compared to what's available today so it's a major major breakthrough and it will democratize access to active safety systems for all budgets

[00:14:36] Wow incredibly cool I'm conscious we're probably making it sound incredibly easy and I'm sure there's been a few learning curves along the way so I've got to ask what were what were some of the key challenges in developing low level sensor fusion and perception technology for advanced driver

[00:14:53] assistance systems and autonomous driving and how did you overcome them I would imagine you've got more than a few stories and tales from your time there but anything you could share yes so so our you know there are many many hurdles

[00:15:10] in developing these complex systems the first thing is you've got to make you know very sophisticated software artificially intelligent software fit into a low-cost processor I mean it can't be into a hundred and fifty or four hundred or five hundred dollar processor otherwise it just doesn't

[00:15:34] scale it cannot fit into the application so you need efficiency you also need to have you know a technology that that actually can evolve over time and I'll explain that in a second because that's very important because as vehicles become

[00:15:54] more software driven there's increasing pressure on OEMs and their suppliers to write deploy and integrate code more quickly and efficient and efficient so it's it's the shift to software-defined vehicles and that is profoundly affecting every function and facet of the automotive industry not only the

[00:16:18] software playing a larger role in core vehicle functions but it also enabling new features such as driver assistance for example as I mentioned but eventually as well autonomous driving OEMs need to incorporate new features that are closer to the start of production they also need the ability to

[00:16:39] quick to quickly and safely push these software software upgrades after the vehicles produced so you can't just have a vehicle that's sitting there for ten years without an update so you know six months in the field of AI is an

[00:16:55] eternity things move too fast so by the time the vehicle is out of production it's out of phase with what is state-of-the-art software so traditionally OEMs would work with a system integrator like a tier one for example to write

[00:17:09] software for each hardware components and then integrate it with code for other parts of the vehicle testing of the integrated software has come late in the process limiting the time available for making additional changes the development of each component and the vehicle platform basically is more or

[00:17:33] less a one-off that starts over and over again it's not scalable it's not practical a new approach is necessary so as a result the OEMs are have shifted from this vertically integrated approach to a more iterative method that incorporates agile software development and collaboration with companies like

[00:17:55] like letter tech with innovative software companies like better tech and they're doing so increasingly with the collaboration of leading SOC vendors so now all of a sudden from something that was defined vertically and everybody

[00:18:10] knew exactly what they needed to do with a specific timeline now you have to have these collaborative multi-layered collaborative approach and one way we have overcome this this this challenge is by you know having our software pre-ported on leading and predominant automotive SOC providers thus

[00:18:33] significantly reducing the risks of development and the risk of adoption as well because now it's a one-to-one direct relationship with the OEM with the SOC and with the tier one integrator so pretty much everybody wins in this

[00:18:50] collaboration and this is a new paradigm and and and it is providing letter tech with a significant advantage over these black box of vertically integrated business models that are out there. And you mentioned that the bright net

[00:19:07] pace and speed of technological change at the moment and the advanced driver assistance systems and autonomous driving market are all expected to grow significantly over the next few months years and beyond so I've got to ask how are you positioning yourself to capitalize on this growth because it

[00:19:27] almost seems inevitable now. Right, this is a very good point I mean you know one of the things that one of the advantages is by being an early mover in this AI based fusion perception software enabled us to build quite a robust and

[00:19:47] growing patent portfolio that is cited by the who's who of the industry so we are recognized for having foundational IP and expertise with solid interiority that is critical for the industry the industry's transition to a software

[00:20:05] defined vehicle so that IP portfolio combined with our expertise and what I would call in a an innovative business model we partner with the customer so the customer is really working hand-in-hand in an open book format where everyone can actually contribute together in the development and that

[00:20:28] deployment of the stack can now be integrated into the OEM's own application development which then it becomes their own so we provide the middle layer they don't have to start from scratch that the SOC provider doesn't have to develop

[00:20:45] software everybody wins so we think that by collaboration historically the tech industry is operated in winner-takes-all and so if your Facebook or if you're Microsoft you want to be the only leader you want to dominate and have 80% market

[00:21:05] share and nobody else well in this AI based you know software for automotive for safety no one can solve that problem on their own you need open collaboration you need ecosystems collaborating or working early together to address this

[00:21:22] problem because like we discuss time to market but also we're dealing with people's lives so if your software is it can be state-of-the-art and can be leveraged by everybody else that's how the automotive industry has operated that's how you get power steering which was invented by Germans anti-lock

[00:21:46] braking systems invented by Japanese they're all standard they're all available to everyone in the automotive industry we think that our fusion and perception layer can be that standard. And another thing that stands out for me is seven years of experience and a substantial patent portfolio in the

[00:22:08] sense of fusion and perception space what kind of advantages does that give you do you think because it seems quite significant again from the the outside looking in. Right well you know a couple of things since we've made significant investments in the development validation of this pioneering

[00:22:29] technology you know this gives us the ability to focus on partnerships and we are engaged with multiple OEMs multiple tier ones multiple SoC providers to really capitalize and make this technology accessible to as many users as possible and we are confident that this business model will accelerate the

[00:22:54] adoption and deployment of AI safety systems in the automotive space and you know at the same time we have built a unique cluster of experts across the globe but mainly in Montreal, Quebec City, Toronto in Canada as well as

[00:23:14] Tel Aviv in Israel where we've got some of the most talented and experienced AI and computer vision scientists that are really continuing to further develop the technology and maintain our industry lead. So there's many advantages one of

[00:23:33] the disadvantage which I will say is sometimes you're so ahead of the market that you have to bring everyone else with you. So we are in the process of slowly and bringing everyone else with us starting with the leading OEMs but it's not so

[00:23:51] much as if there was a clear roadmap everybody knew exactly it existed and people could just follow we just we have to continue to educate not only the automotive industry but also you know potential individuals users of these

[00:24:10] technologies and events investors as well because this is so new that people aren't familiar with it. And I think another aspect of every single tech project in every industry now is this there's almost a huge focus on things

[00:24:27] like business value and ROI from every tech project. So from that side of things how does Ledatech software solution help maybe reduce costs for OEMs and tier one and two automotive supply as well delivering improved performance and

[00:24:43] scalability I appreciate it must be quite a delicate balance but is that something that you do too? Absolutely well the way to think about it is that you know as I mentioned earlier cars are becoming more and more complex they're

[00:25:01] outfitted with multiple sensors and processors but how they process the data from these sensors matters. Object level fusion systems today which are currently used in most cars they interpret each sensor independently and often have troubles you know in different harsh environments different environments

[00:25:23] so our pioneering AI based fusion perception software really enables more accurate view of the environment by taking existing hardware existing cameras existing radars existing sensors and doubling their performance natively by software. This basically enables you to do more imagine you know I would say you

[00:25:51] know you have a car its top speed is 150 kilometers per hour I add my software all of a sudden you can do 300 kilometers per hour you know what that would do right and that sort of performance is a scaling of our software

[00:26:10] is quite it's quite revolutionary. So this allows you to as a car maker or as a designer actually use lower grade hardware to do different combinations of sensors this gives tremendous flexibility and we believe that this is

[00:26:31] the way to go. Incredibly exciting and speaking of exciting I've got to ask what was the significance of going public on NASDAQ's? It feels incredibly exciting how's that impacted your growth and strategic initiatives it feels like

[00:26:46] there's a lot going on there but what kind of impact does that have? Yeah I mean operationally this has not changed much for Ledertech to be fair you know we have been operating since we've had that tier one investors global investors

[00:27:02] we've had to report on a quarterly basis to our board to our investors. Going public provides a different access to capital it gives us more use or more vehicles to raise capital. We're like a biotech company you know with we raise

[00:27:22] capital for the next milestone and we are in this next phase next next milestone the next milestone is we expect one of the many OEMs that we are engaged with or many tier ones or partners to convert into a significant

[00:27:40] win and that can translate into hundreds of millions of dollars in revenue so being public gives us access to capital to go to the next phase easier access to capital. Then the other thing too is that we're dealing with very large companies

[00:27:59] traditional in many ways the European, US or German manufacturers or even Koreans and Japanese they're traditional in their ways and it's a long long cycle with a high barrier to entry so they want to know that you have

[00:28:21] the stamina and the capital structure in place to be able to support them for the long term and so being public gives them that assurance that we will be able to continue to raise capital as long as we continue to progress with them. And now

[00:28:37] you are public I'm not sure how much you will be able to share with me today but I'd love to get some teasers out of you as we look ahead what are Ledertech's key priorities and goals for the future particularly in advancing some of the

[00:28:50] technology around autonomous driving etc that we've discussed today. Any teasers or anything you can share around that road ahead? Yeah I mean first things first is you should expect us think of us like Microsoft OS for example where we will be releasing versions of our software that are

[00:29:11] increasingly being more complex more sophisticated. For those of you who are familiar with chat GPT the difference between chat GPT 3 and chat GPT 3.5 was significant. The difference between chat GPT 3.5 and chat GPT 3.4 was just out of

[00:29:31] this world and so you should expect that our evolution of our software will be step functions better with every release. So we expect to do a couple of releases before the end of this year and more next year. Combined with that we

[00:29:49] expect to continue to expand our partners. We have three SoC partners that are being announced this year. That's Texas Instruments, that's Arm and that's Black Sesame. We are working with another leading SoC provider that covers the

[00:30:08] China market with Black Sesame and that leading SoC provider which we're working on should give us 90% coverage of the China ADAS market. So we are very very strongly positioned there and so you should expect more announcements along

[00:30:25] those lines and we are also working with tier ones that are you know building products with our software. One of which we've announced for a couple of years now is Chicosa out of Spain. They're a leader in cameras for parking

[00:30:44] applications and they're expanding their business really beyond and they're growing you know at a very fast rate. The next generation solutions will be based on our software and so we but we expect to announce more tier one partnerships

[00:31:00] and then eventually over the next six to twelve months we're working with almost you know a dozen type of OEM projects. One of which these OEM projects will convert into a significant design win which will then be a step

[00:31:20] function for Laratech in terms of revenue but also in terms of value creation. So lots going on. We'll continue to scale and grow the company and then build our customer rosters. Wow such an exciting space to be in right now and I feel that this story is

[00:31:39] going to continue to evolve for the rest of 2024 and beyond. So I'd love to get you back on next year see how things are progressing but as a little thank you for shining a light on this and demystifying it and put it in a language

[00:31:53] everyone can understand today. I'm gonna see if there's something we can do for you now because some of the biggest names in business VC funding and tech have either been guests or maybe listened to this podcast. So is that a person you'd

[00:32:05] love to have a private breakfast or lunch with? Who would it be and why? He or she might just get to hear this let's see what we can manifest but who would it be? You know since we operate in the same industry this man has been in

[00:32:21] the news for over the past couple of decades. His name is Elon Musk. I think you know he gets a lot of stick for saying what crosses his mind but he's got a brilliant mind and would love to be able to hear his perspective and hear see how

[00:32:40] he sees the world in a one-to-one basis. I thought that would be fast I think that would be fascinating to hear him and without the distraction to unfiltered connection. I should have known you would have gone Elon Musk

[00:32:57] straight away a great choice there. We'll throw that out into the ether let's see what the universe can deliver there but for everybody listening at home maybe they want to find out more information about Lettertech dig a little bit deeper

[00:33:11] on the topics that we discussed today where would you like to point everyone listening? Well you know you can you can follow us at letter on our website lettertech.com but you can also follow us via the Nasdaq to a ticker L DTC but

[00:33:28] the most efficient way is usually through LinkedIn we are very active as a company on LinkedIn on this platform and so you can always stay tuned either with one of these three platforms lettertech.com on the Nasdaq L DTC or on

[00:33:49] LinkedIn. Well I've loved chatting with you today especially because it's such an exciting topic at the moment because advanced driver assistance systems ADAS and autonomous driving or AD for short is the largest market within automotive

[00:34:05] software right now it's expected to grow at I think 11% CAGR to 42 billion by 2030 and that's not that far away so incredibly exciting I'd love to have demystified it all today and like as I said a few moments ago I'd love to get

[00:34:21] you back on early next year see how things are progressing but cannot thank you enough for investing your time with me today and demystifying this thank you. Thank you for having me Neil. As we conclude today's discussion I think the

[00:34:34] revolutionary strides that Lettertech is making in the realm of autonomous driving technologies it's clear that the road ahead is both exciting and challenging and their approach to integrating AI to enhance vehicle perception capabilities it doesn't just promise safer roads but also opens up

[00:34:53] new avenues for innovation in automotive design and functionality so I hope you enjoyed today's episode as much as I did but I'd love to ask you the questions now not just my guests what are your thoughts on the future of

[00:35:05] autonomous driving technologies how do you see AI shaping the evolution in that vehicle that you drive every day please email me now tech blog writer outlook.com Twitter LinkedIn Instagram at Neil C Hughes but that is it for today I've got

[00:35:21] another guest another topic lined up for tomorrow I cordially invite you to join me again but thank you for listening today and until next time don't be a stranger