Automotive Electronics Predictions
As we are, approaching the close of 2020, undoubtedly the most unusual year for automotive electronics. 2020 was the year that most annual plans got thrown out the 1st quarter, and the rest of the year was felt like guess work. In March we experienced the most dramatic pull back, followed my an onslaught of Zoom, Team, Webex, Skype, Slack calls that spanned more hours than the typical work weeks. Many of us were zooming daily with 3 different continents in our make shift home offices wearing our pajamas.
That said with the promise of some future normalcy in 2021, I thought it would be good to pull out the crystal ball and make a few predictions to usher in the New Year.
- Autonomous Trucking will pave the way for Autonomous Passenger vehicles. We will see long-haul trucking from depot to depot, highway only driving, on special corridors that are enhanced with special 5G connected land based sensors arrays, and emergency remote driver intervention (drone mode). With a simplified operational design domain it will be easier to develop a system. Highway only mode will reduce the complexities. A drone mode will provide another layer of safety on top of the redundant compute. The trucking industry can bear the higher system costs of advanced sensors (cameras, lidars, radars), the redundant central compute/communication systems, and the cooling systems.
- Xilinx devices: will be power both the advanced sensors, central compute in the trucks, 5G communications and the cloud based data center support systems.
2. Know thy self. Vehicle that can heal themselves. Well maybe not exactly fix themselves but know when to go in for a repair before the car damage is more serious. Vehicles will have prognostics, the ability to "self-diagnosis", virtually eliminating the worrisome situation of being stranded when your vehicle breaks down. Car already have an array of sensors to monitor engine temperature, air flow, shaft rotation, spark timing, etc. we can now combine these with prognostics, not to mention having a "microphone" to listen for the signature failure "tells" of squeaks, and thumps.
- Xilinx devices: will be provide the AI capabilities either on the vehicle (edge) or in the data center (cloud)
3. Student Driver mode. Just as a AI can create feedback as to your driving habits on fuel efficiency and automated vehicle equipped with cameras radars, and lidars can provide a safe environment to a new drive to insure accident avoidance but also coach in real-time just as a parent would teach their children. Not only can we use the cameras for ADAS you can use this as a video playback feature. "lets go to the tape, the tape doesn't lie", no longer do we need to argue with the student driver we have the video data to provide proof to provide feedback.
- Xilinx devices: once again will power the AI to provide the necessary feedback lessons.
4. Vehicle Valet; Knight Rider had "Kitt", Space Odyssey had "HAL 9000". Ironman had "Jarvis". Vehicle valet will keep you safe, informed, entertained, and productive. Voice, and Gesture UI will intuitively understand your commands. Augmented reality will aid in navigation, providing points of interest or alerting of safety issues (Using VR projection)
AI valets will use a unique blending of AI that combines commands: Voice, Gesture, with contextual awareness inferred from Body Pose, Eye Gaze. These AI technologies are only in their infancy and the HW systems will need to adapt as they AI techniques improve. Adaptive HW will maximize the longevity of the system via OTA updates.
Valets systems will be able to: i) anticipate your driving mood, leisurely Sunday drive, or get there in a jiffy. ii) Leverage complex sensor arrays (Camera, Radar, Lidar) to provide VR overlays
- Xilinx device support the AI, and used in variety of intelligence sensors, but more importantly they will provide the adaptability via OTA Hardware, allowing you to update your AI to the state of the art.
5. WFV = Work From Vehicle. Your car is your remote working hub. As the pandemic has proved worker productivity can be achieve from WFH (work from home), but why limit yourself to your home. Your vehicle can be your work hub, work as you go about your daily lives, get your next video conference call from your vehicle.
- Xilinx device will support various image signal processing (ISP) for cameras utilized for DMS/OMS but those same ISPs are not intended for human vision, Xilinx devices can support both ISPs, and optimize them to maximize the image for human vision or machine vison.
6. EV growth will transform urban and rural planning. Like train stations in many countries like Japan, China, and Germany. Train stations are a hub for shopping and restaurants. EV will necessitate the need for charging stations to which consumers will shop and eat as they wait.
- Xilinx devices: will be impactful curbing the range anxiety with better EV motor drive, and will also help accelerate the charging process as SiC mosfets start to phase in.
7. Battery recycling will emerge as a business to deal with disposal and recycling to prevent future environmental impact. Why stop at limiting carbon emissions from internal combustion engines (ICE)? Batteries are also an issue for the environment. A new industry will be necessary to jump in to protect the environment, but also due to the scarcity of rare earth minerals such as Lithium, and Cobalt used to produce batteries to power the EVs
- Xilinx device - will do its part by supporting predictive AI algorithms to help maximize the battery life by monitoring the state, optimal charging, and discharging
8. Automotive OEM will unlock reoccurring revenue streams. Car companies have long aspired to use the personal owned vehicle (POV) as a platform for recurring revenue. One example of this was the foray into satellite radio with XM Satellite radio and Sirius Satellite radio. Another example, GM's OnStar provides vehicle security, emergency services and turn-by-turn navigation. Both these examples prove that subscription services can be successful in the automotive market. One of the lessons learned from is that the hardware needs to be adaptable platform that can support "features" yet to be designed today, and have long service periods (vehicle life). This will mean vehicles will have "more" hardware than necessary, as they create upgradeable HPCC (high performance Central Computing) nodes in the vehicle. They will unlock the ability to sell "features" like blind-spot detection, trailer hitch, similar to the way mobile phone sell "apps". These can be sold to the original owner or even a subsequent owner of the vehicle.
- Xilinx device are innovation platforms, that support OTA hardware. A unique capability that can update both software and hardware.
9. OEMs will create brand loyalty programs by extend the Vehicle "apps" where you can take the features purchased in a previous vehicle to a new vehicle so long as you pay your "subscription fees", the same why can move apps you have purchased on your mobile phone to your new phone. Some of these "apps" will be created by the OEM, Tier1s or even 3rd party developers. This will spawn a new ecosystem of IP developers for automotive "apps" (features)
- Xilinx devices have advanced security features to support the protection of critical IP.
10. Vehicle networking will go wireless. As with the home, we are becoming more advanced in wireless connectivity. Vehicle cost is heavily influenced by the cost of wires harness and connectors, and the weight of these components affect the range of the vehicle. Eventually, wireless networks will have a FuSa strategy, that will be fast, reliable, and low latency (signal delay). Yes this is probably a ways out in time, but with all the advancements in technology it is bound to happen at some point.
- Xilinx devices have a long history of serving the communications market, and have enable software defined radios via Dynamic Function eXchange. DFX is yet another unique capability of Xilinx devices, that allows you to repurpose the field programmable real-time while active to support mutually exclusive functions, or sequential operations, ideal for a automotive communications gateway.
Well those are my predictions. If you are interested in learning more about Xilinx in automotive, sign up for Xilinx Adapt Automotive: Drive to the Future.
