Today’s connected and autonomous vehicles pack many electronics and computing power to perform Driving Assistance functions. Those devices not only consume a lot of energy but also add substantial weight to the car.
In 2005, “Standley,” the winner of the original Darpa challenge for autonomous vehicles, featured five roof-mounted Sick AG LIDAR units, an internal guidance system utilizing gyroscopes and accelerometers, a video camera used to observe driving conditions out to eighty meters (beyond the range of the LIDAR), and six Intel Pentium M based computers in the trunk. All that equipment allowed Standley to self-navigate a predetermined 150 mile route across the Mojave desert.
A modified version of the European Volkswagen Touareg, the car had a diesel engine. If it were an electric car equipped with today’s batteries, it wouldn’t have been able to drive for more than a few miles; the onboard systems would have drained the batteries in a few minutes.
Today, 17 years later, a connected vehicle with Advanced Driver-Assistance System (ADAS) Level 2, performing a few autonomous tasks such as lane-changing warning, emergency braking, and parking assistance, has 100x the processing power and the sensor capabilities, and uses just a fraction of the power.
The arrival of full-electric vehicles poses an additional challenge for the electronics industry. Unlike their internal combustion engine counterparts, EVs use the energy stored in their batteries for everything, and onboard electronics – besides the electric motor – consume a substantial amount of energy that which, in turn, reduces the EV range.
Furthermore, automakers are constantly improving their EVs with more features, including more ADAS features, infotainment equipment, and additional comfort, all of which, again, increases power consumption.
Connectivity alone is one of the significant power drains. A connected car, without autonomous features, could be sending and receiving several gigabytes of data daily using cellular networks. It is estimated that a basic ADAS level 1 car could exchange over 100 terabytes in less than five years, mostly to update maps and send sensor data. Using 5G networks and C-V2X communication, a fully autonomous vehicle will transmit over five terabytes in 24 hours of operation.
Reducing vehicle capabilities is not an option
Investment in smaller electronics and low-power computing is necessary to reduce the power needed. Additionally, efficient edge computing should reduce the need to upload massive amounts of data to the cloud, reducing the use of power-hungry cellular connectivity.
Today’s connected electric vehicles feature an army of sensors to understand the environment and additional communications to link to other vehicles (V2V) and the road infrastructure (V2X). Even the most basic models have proximity sensors to help drivers park and move around tight spaces. Advanced models also have several cameras to detect other vehicles and infrastructure around them, and cars with sophisticated ADAS features also have other sensors such as Radar and LiDAR.
Fortunately, those electronic subsystems have improved much faster than the vehicle itself. A current LiDAR unit from Velodyne can capture several million data points per second, creating a precise image of the road ahead. In comparison to the one used for the Darpa challenge, a Velarray H800 can fit neatly behind the windshield of a truck, bus, or car and use less than 5% of the power.
Onboard electronics and processors are advancing at computer speed
Companies such as Nvidia, Arm, NXP, Infineon, Qualcomm, and many others are heavily invested in the automotive industry. As cars become more like computers on wheels, optimizing and shrinking their CPUs while enhancing the performance is necessary to reduce power consumption.
Recently, Arm unveiled its Scalable Open Architecture for Embedded Edge (SOAFEE) architecture, a heterogeneous computing environment optimized for automotive within a power performance window required for ADAS level 4 and level 5, all within an SoC that requires little or no external cooling. Last year, Nvidia announced their Atlan Autonomous Vehicle Platform, a next-generation SoC to deliver more than 1,000 trillion operations per second (TOPS) on a single chip.
These are just a couple of examples of how fast computer processing advances in the automotive industry. This new level of computer power, which could easily fit under the dashboard, enables onboard processing of all the vehicle sensors and other subsystems, reducing the need to upload information to the cloud.
Current and future autonomous vehicles would be able to operate without any external assistance, except for their local communication with other vehicles and road infrastructure.
DC Charging could further reduce onboard electronics and enable fast charging
For people charging their electric cars at home, the most popular option is a wall-based AC charging station. Unfortunately, plugging an electric vehicle into an AC charging station means that the necessary AC to DC conversion to charge the batteries happens inside the vehicle.
An electric vehicle needs to connect to a fast-charging station using DC power to enable fast charging. Today, the infrastructure for fast charging stations, the ones now found on businesses’ parking lots, city streets, highways, and some traditional gas stations, feature different connectors, including AC and DC power.
It would be a real game-changer if the manufacturers of electric vehicles could remove the AC transformers, and EV owners could also use DC charging at home.
Companies like Wallbox are now selling a home solution for DC charging. Using a CHAdeMO connector, the Wallbox’s home Quasar 2 charger features DC charging up to 11.5kW, allowing car’s battery to power the house, even during a power outage. For example, a fully charged electric car with 75 kWh capacity can supply electricity for a typical home’s essential needs for over a week. To learn more about EV charging, read How OBCs, fast DC chargers overcome range anxiety in EVs.
Electric vehicle manufacturers need to start thinking like computer companies
Internal combustion engine vehicles are facing phase-outs around the world. EV manufacturers need to work closely with the technology companies making onboard electronics, cellular communications infrastructure, charging stations, and cloud services to take advantage of this trend.
Optimizing existing electronics is not enough. Automakers need to adapt their production lines for flexibility of electronic design. It would help them take advantage of newer technologies quickly and better manage the shortages in their supply chain when some components are difficult to procure. Today’s electric vehicle is more a computer and batteries on wheels than a pure transportation system. Automakers need to think outside the box and realize this unstoppable reality.
