Good afternoon, everyone. I'm Raj. I'm excited to be here today to talk about the fascinating and transformative field in the automotive industry, which is advances driver assistance systems and autonomous driving. Let's dive into it right away. In this presentation, we'll delve into the exciting world of adas and autonomous driving. Think of adas as your car's superpower, which enhances safety and makes driving lot easier. We'll explore the key components that power these technologies, the impressive functions they perform, and how we are steering towards fully autonomous vehicles. Beyond this tech talk, it's all about making our roads safer and transportation more efficient. So let's navigate through the advancements that is shaping the future of driving. I'll talk about the core components and functions of adas, the computing needs of adas, and finally touch upon the evolution towards the autonomous vehicles and the impact on road safety and transportation. So what is adas and ad? ADAS stands for advanced driver assistance systems. It's a suite of technologies designed to enhance vehicle safety and driving experience. From collision avoidance systems to adaptive cruise control, ADAS represents a digital copilot seamlessly integrating technology with driving. And when it comes to autonomous driving, actually the vehicles they navigate and operate without human input. This is a shift from driver assist features to fully autonomous vehicles, revolutionizing the concept of transportation. Now, let's take a look at the brains and eyes behind the magic of adas. Imagine your car being equipped with superhero squad, each member playing a crucial role. There are several members here, like radars, cameras, ultrasonic sensors and lidars. Let's see the functions of each of these members. Consider a long range a radar. It scans the road ahead, detect the objects from a distance, helping the car to anticipate and react to potential dangers long before they become an issue. And then short range radars. They have a keen sense of proximity. It focuses on the immediate surroundings, ensuring your car navigates through tight spaces, handles parking, and keeps an eye on blind spots. And cameras. Interpret traffic signs. Recognizes lane markings and provide a real time visual feed, offering eCarx a comprehensive view of the environment. And there are infrared cameras which act like stealthy night vision experts. These cameras see what the human eye can't in low light conditions, ensuring safety even during night when there is not so much light. And there are ultrasonic sensors like these are like silent guardians. They are the ones responsible for detecting obstacles in close quarters, making sure your car doesn't inadvertently bump into anything during low speed manures. Last but not the least, lidar, which is the precision mapper using laser beams, Lidar creates a detailed 3d navigation of surroundings, allowing the car to navigate complex environments and understand the world with incredible accuracy. Together, these components form the powerhouse of adas, working seamlessly to make the driving experience safer and smarter. It's like having a high tech team ensuring you're in good hands every time you hit the road. Here are a few examples of advanced driver assistance systems. There's some ADAS features, basically. So for example, front collision warning. It alerts the driver if a potential collision with the vehicle ahead is detected. And there is automatic emergency braking which automatically applies the brakes if the driver doesn't respond in time. And for example, there is a lane departure warning which notifies the driver if the vehicle unintentionally drifts out of its lane. And there is lane keeping assist which can gently steer the vehicle back into its lane to prevent unintentional lane departure. So you see a number of ADAS functions which are divided into safety package, in comfort package and in parking package. So depending on which kinds of features the car makers choose, you will have those functions in mid, premium or even entry kind of vehicles. And maybe just to explain a couple of functions more, there is for example adaptive cruise control. This maintains the speed, but can automatically adjust the speed based on the distance to the vehicles ahead. It helps in maintaining a safe following distance in varying traffic conditions. And there is blind spot detection, which alerts the driver if there's a vehicle in the blind spot. Typically with a visual or audible warning. It enhances and it gives an awareness during lane changes. And then there are park assist technologies and automated parking assistant which basically it steers the vehicle into a parking space, providing guidance to the drivers. And there's traffic sign recognition which uses cameras to identify and interpret traffic signs displaying relevant information on the dashboard. It helps drivers stay informed about speed limits, stop signs and other road signs. And there is something like cross traffic alert which warns the drivers of approaching traffic from sides, especially when backing out of parking spaces. It advances safety in situations where visibility may be limited. These examples, they showcase the diversity of ADAs features, contributing to safer and more convenient driving experiences. Coming to the compute power which is required for ADAs and Ad, I would like to discuss this compute in three dimensions. The first of it is cpu or the scalar computing. The second one is gpu or for both three d and parallel capability. And certainly the most important for ADAs and Ad which is the computing for machine learning and inferencing offload. So let's have a look at cpu load first so today we see socs that reach something beyond 100 kd mips. There are x 86 socs from multiple vendors which they offer like something about over 500 ktmips and beyond we have server socs which can deliver as much as 5000 ktmips. We can see quite a significant dynamic range here. What's interesting is when we start looking at ADAs and Ad, we see a higher level of computational capacity for cpu, as much as 1000 kd mips in the case of robotaxi class of applications. On the gpu front, let's see what options do we have silicon wise, we see arm igpus which are the integrated gpus from half to about five t flops. We see discrete gpus for pc, mobile and desktop, ranging from 50 t flops and then data center discrete gpus providing even higher capability of t flops and other computational capacity. So what does it look from an ADAS perspective? Even in ADAS applications, we see a demand of gpu computing. However, this isn't as much. Gpu requirements are ranging from 0.5 t flops, maybe something like to one t flop. When we go to machine learning, we see a pretty wide range of capabilities. These are socs from five to 200 intake tops, depending on the vendor in question. Data center is offering like 500 intake tops and beyond. This is what is available in the market today. Obviously, the market is responding rapidly and aggressively to the increase in the capabilities for ADas and ad. We need as much as something like 1000 intake t flops, a very high level of dynamic range. We can also combine two or more socs to get the required amount of tops depending on the application. So with such high compute socs at our disposal, and more and more autograde socs which are available in the industry, we are able to address the needs of all autonomous vehicles, especially from l three to l five autonomy vehicles. Now let's see what kind of chips are available with some examples here I got some examples of chips which we use for adas on autonomous driving. So, cutting edge chipsets like this play a pivotal role in meeting the computational demands, which is essential for real time decision making and control. As an example, horizon robotics. They have an emphasis on edge computing. With them, they excel in processing sensor data directly on chip, particularly in computer vision and object recognition, and then Nvidia or in and Thor. Of course, the architecture stands out for its powerful gpus, enabling high performance parallel processing and simultaneous support for multiple neural networks, which contributes to robust perception capabilities in varied driving scenarios. And then Texas instruments like TitDA four series. They focus on sensor fusion, efficiently integrating data from diverse sensors to provide a comprehensive understanding of the vehicle's environment and then mobili. For example, they specialize in vision based solutions with its iq chips, excelling in computer vision algorithms for applications like lane departure, warning and collision avoidance. So these chipsets, they collectively address the intricate needs of adas and autonomous driving. Their optimized designs support tasks such as sensor fusion, real time perception, and decision making, which is crucial for enhancing both safety and autonomous capabilities. Now let's have a look into the software elements. They can be broadly classified into, like fusion, perception, planning, middleware, and components for safety. So what are these components? The fusion software component. It acts like a brain of adults, merging data from diverse sensors such as cameras, radars, lidars, and create a comprehensive and accurate representation of vehicle surroundings. Perception software, on the other hand, involves advanced algorithms that interpret sensor data, recognizing objects, pedestrians, and road signs. This components is crucial for real time decision making, providing the system with the ability to identify and respond to dynamic changes in environment. Middleware, with platforms like QNX and Autosar, serves as the communication bridge between various hardware and software components in the ADAS ecosystem. So, QNX, it's a real time operating system often used in automotive, which ensures consistent and predictable performance. And Autosar, it provides a standardized framework enabling seamless collaboration between the different software modules which we have in the system. So also we need to ensure that these ADAS software components, they adhere to rigorous safety standards, and this is usually done by the NCAP certification, through the NCAP certification. So functional safety, which encompasses standards like ISO 26262, is paramount in the development of ADAS software, which emphasizing the implementation of measures to prevent or control to control system failures that could compromise safety. So these elements collectively form the robust software backbone for adAs. Now, in the landscape of autonomous driving, there are five levels of autonomy, from l zero to l five, each representing a distinct step towards vehicles that can navigate the roads without human intervention. So at level zero, we have no automation where the driver is fully in control. As we ascend through the levels, we witness the gradual integration of automation, for example from basic driver assistance features like some warning features at level one, to driving features like adaptive cruise control in level two, to conditional automation in level three, where the vehicles can handle certain scenarios independently. So reaching level five, the pinnacle of autonomy. We envision fully self driving vehicles that can operate in any environment without human input. So this evolution reflects not just a technological progress, but also a paradigm shift in how we perceive and experience transportation. So before I close. Let's have a look at the impact of Adas and ad on road safety and transportation. ADAs emerge as vigilant copilots, significantly enhancing road safety through features like collision avoidance systems, lane departure warnings, and adaptive cruise control. Simultaneously, autonomous driving has a revolutionary impact on safety, where vehicles are capable of swift, split second decision making, reducing accidents, and enhancing traffic flow. It also enhances accessibility and inclusivity in transportation, opening new avenues for individuals with disabilities and accommodating diverse needs. The integration of these technologies brings many environmental benefits to optimizing traffic patterns and reducing fuel consumption and emissions, thereby contributing to a greener and more sustainable transportation ecosystem. So, in conclusion, the combined influence of ADAs and AD marks a revolutionary shift in the future of transportation, from advanced safety features to redefining accessibility and promoting environmental sustainability. I hope the talk was end of the presentation. I hope this short talk about ADAs and ad has sparked interest in some of you and helps you to further deep dive on some of these topics. Thank you. Have a nice day.