Power Electronics in Self-Driving Cars
Authors: Akshay Joshi, Soham Kamble, Atharva Karwande, Tejas Kolhe, Pranesh Kulkarni.
Automated vehicles have drawn increasing attention in recent years. These need sensors and computer processing which will understand the surrounding environment and make real-time decisions. However, these vehicles aren’t fully automated, and to reach higher levels of automation, a lot of sensors and systems should be implemented to control the vehicle in all real-world circumstances. The addition of advanced driver assistance systems (ADAS) to a vehicle is a task in itself. A vehicle has a limited area for sensors, wiring, power supplies, and computer processors. Additionally, all these new elements added to make a vehicle automated, consume power. While individual sensors would possibly not be large loads, the power drawn by a large number of sensors can compound to be significant. These further electrical loads expand the auxiliary load profile, thus reducing the range of an automated electric vehicle compared to a standard electric vehicle.
Most new studies regarding automated vehicle systems use an electric vehicle (EV) rather than an internal combustion engine (ICE) vehicle. EVs are inherently easier to regulate using automated driving sensors and systems because control is accomplished electrically rather than mechanically. Furthermore, EVs have fewer moving elements than ICE vehicles, which might lead to improved reliability. In addition, an optimized EV powertrain in a fully automated transport system only requires one-third of the energy of an equivalent ICE vehicle.
With the increasing demand for environmentally friendlier and higher fuel economy vehicles, automotive companies are specializing in electric vehicles, hybrid electric vehicles (HEVs), plug-in hybrid electric vehicles (PHEVs), and fuel-cell vehicles. These vehicles would also enable meeting the demands for electrical power because of the increasing use of electronic features to boost vehicle performance, fuel economy, emissions, passenger comfort, and safety. In electric vehicles, HEVs, PHEVs, and fuel-cell vehicles, the challenges are to achieve high efficiency, ruggedness, small sizes, and low prices in power converters and electric machines, as well as in associated electronics. In particular, in fuel-cell vehicles, a power-conditioning unit such as a dc-dc converter for matching the fuel-cell voltage with the battery pack may additionally be necessary. Furthermore, the combination of actuators with power electronics not solely improves the overall system reliability but also reduces the cost, size, etc. in addition to power electronics; the technology of the electric motor plays a major role in the vehicle’s dynamics and therefore the type of power converter for controlling the vehicle in operating characteristics. Power electronics is an enabling technology for the development of these environmentally friendlier vehicles and implementing advanced electrical architectures to fulfill the stress for increased electric loads.
For a self-driving car, light detection and ranging (lidar) is a versatile light-based remote sensing technology that recently has been the topic of great attention. In its known form, a lidar bounces a laser beam off an object or a target and uses the reflection to see a number of its properties. Tremendous advances in autonomous mobile machines, as well as robots, drones, and automobiles, have led to a necessity for real-time, correct 3-D mapping for navigation and collision avoidance. For an autonomous driving computer, lidar is that the ideal technique to acquire data. Lidar provides the foremost strong and data-efficient means that of assembling high-resolution, high-accuracy distance data. For power electronics and lidar, the primary relationship revolves around the light and its driver.
Switched Mode Power Supply (SMPS) in self-driving cars:
Switch Mode Power Supplies or SMPS have replaced the traditional ac-to-dc power supplies as a way to cut power consumption, reduce heat dissipation. This concept contributes to reducing the weight of the vehicle. Basically, it is the type of power supply that uses semiconductor switching techniques, rather than standard linear methods to provide the required output voltage. This technology is based on power electronics devices which can give higher efficiency in power switching applications. The pulse width modulation technique is used to turn off and turn on the semiconductor devices. These power electronics-based devices having small sizes can operate under high switching frequencies so they are used for high-frequency switching.
SMPS Applications in the Power Train System of self-driving cars:
1. On-board Charger (AC to DC Converter):
Self-driving cars consist of hundreds of sensors, microprocessors, and electronic components, which work on batteries that need to be charged. So for charging these batteries, the AC power supply has to be converted into DC because the power can be stored in the batteries in the form of DC only. AC to DC conversions is done by power electronics converters called rectifiers. A rectifier is a power electronics device that converts AC, which periodically reverses direction, to DC which flows in only one direction.
2. Traction Inverter (DC to AC Converter):
Most autonomous driving cars use electrical power rather than traditional engines which use fuels like petrol and diesel. Tesla which is the leading industry in producing autonomous driving cars built electric vehicles only. Electrical motors are generally used in electric cars to convert electrical energy to mechanical energy. For this purpose primarily DC motors are used but to increase reliability and efficiency, AC motors are used. Many power electronics components are used for developing controllers for AC motors. To supply the power to the AC motor, the power stored in the battery which is in DC form is used by converting from DC to AC. For the conversion of DC to AC, power electronics devices such as DC to AC converters (inverters) are used.
3. DC to DC Converters:
The self-driving car consists of many electrical components which operate in a wide range of voltages. Devices like display monitors, microprocessors, sensors, air conditioning systems, Lidar systems operate in different voltages. So these devices run on a power supply from batteries. But we cannot directly provide the DC supply coming from the battery. There should be a device that converts a source of direct current (DC) from one voltage level to another. So power electronics devices like DC choppers fulfill these requirements which have a wide range of conversion ratios. In these devices, the output dc voltage can be varied by using pulse width modulation control.
Power electronics in LIDAR applications:
Self-Driving Cars, which were a dream just a few years ago, are becoming reality. This type of advancement is made possible due to high-end performance in some smart sensors. ADAS (advanced driving assistance systems) such as lane detection, pedestrian detection, and adaptive cruise control, ensure the safety of the driver and keep up with the safety standards. One of the most significant components of the ADAS system is LiDAR. It helps in general to detect and map all surrounding elements surrounding the vehicle.
LiDAR stands for Light Detection and Ranging. It is detection and measurement by electromagnetic radiation in the optical band. It is very similar to radar, but it works on light instead of sound waves. Compared to RADAR, the optical device has higher resolution even at long distances and thus helps in 3D mapping of surrounding with increased accuracy to avoid a collision.
To detect or measure, the sensor beams out pulses of light in the environment and the photosensors receive the reflected light for the object. And the time duration between the emission and receiving of the light gives us an estimate of the distance and the speed of the object.
But a very important aspect of these criteria to be met is the pulse width of the LiDAR. The narrow pulse allows reflection to be separated and the distance between targets to be resolved. A wide pulse causes the reflected pulse to overlap, and the objects cannot be distinguished. Ideally, the pulse width must be in the order of 0.1 nanoseconds and the frequency must be in the order of THz.
Hence the ultimate performance of the LiDAR is limited to the pulse width. A shorter pulse width means greater resolution and fast measurement and less computation.
There are many sources used in the LiDAR. The most commonly used is LED due to its small size and low cost. Although it seems simple enough, to turn on and off a diode to generate a pulse. But the driving circuit operates at high currents in the order of 10s of Ampere. With this large current flowing through the circuit, there is a considerable rise in the temperature of the devices and sometimes thermal runoff happens in these devices.
We can certainly find a Power MOSFET that can meet these high voltage demands and temperature, but it is impossible to switch it at a rate of 1ns. A very fast Si MOSFET is capable to switch at a rate of few nanoseconds, but it will not extract the full benefits of the laser. Hence, due to the limitations of the devices, there is a certain decrease in the performance of the system.
But, with the advent of the Gallium Nitride (GaN) FETs, there is a dramatic change in the way the LiDAR with a laser works. It allows a very fast switching speed and can be generated with a size suitable to the SOC requirements. GaN FETS have a higher mobility speed of the carriers. GaN MOSFETs have lower conduction losses, and better thermal performances, and smaller sizes and costs. All these features satisfy the needs of the driver circuit switch component.
The use of GaN components in commercial devices is only at the beginning. The technological solutions, considered to be impossible or too complicated a few years ago, are proving to be successful in many areas, such as in power drivers for LiDAR systems. Thus, confirming that, in the coming years, the field of power electronics will be dominated by WBG devices, which can solve the technological limits of the “old” semiconductors devices.
The solution for reducing EMI/EMC Emissions:
Automated EV is incomplete without the sensors used. These sensors work on radio frequencies that get disrupted because of interference from EMI/EMC from different devices.
EMI stands for Electromagnetic Interface. It is the disruption of the operation of an electronic device by an electromagnetic field. This is basically only created by moderate or high-powered wireless transmitters. For example, it is likely that a user has experience EMI if living near a radio or television transmitter.
EMC stands for Electromagnetic compatibility. EMC of a device is the ability not to be affected by electromagnetic field i.e. EMF and not to affect other systems operation with its EMF when it is operating in an electromagnetic environment. EMC represents electromagnetic emission, susceptibility, Immunity, and coupling issues.
In recent days for optimum functioning and reliability, automobiles have more electronic components and systems. When we look at the design of an electric vehicle, we notice a lot of electrical and electronic systems packed into a little space. Electromagnetic interference, or cross talk, occurs between various systems as a result of this. These systems could malfunction or even fail to operate if EMC is not properly managed.
The LT8650S is a dual output monolithic synchronous buck converter with a high input voltage capability and low EMI/EMC emissions. Its input voltage range of 3 V to 42 V makes it excellent for automotive applications, such as ADAS, that must regulate through the cold crank and stop-start scenarios with minimum input voltages as low as 3 V and load dump transients exceeding 40 V. The switching frequency of the LT8650S can be configured and synchronized anywhere between 300 kHz and 3 MHz with a 2 MHz switching frequency and a 40 ns minimum on-time, it can perform 16 VIN to 2.0 VOUT step-down conversions on the high voltage channels. The switching frequency of the LT8645S can be configured and synchronized between 200 kHz and 2.2 MHz The LT8645’s design drastically minimizes EMI/EMC emissions by combining very well controlled switching edges with an internal architecture with an inbuilt ground plane and the usage of copper pillars instead of bond wires. Internal top and bottom high-efficiency power switches, as well as the required boost diode, oscillator, control, and logic circuits, are all integrated into a single chip in the LT8645S.
The industry trend of power electronics in self-driving cars:
According to a report published by Allied Market Research by the title “Power Electronics for Electric Vehicle Market by Application and End-Use: Global Opportunity Analysis and Industry Forecast, 2019–2026” the market for power electronics for electric vehicles were valued at $2.59 billion in 2018 and is projected to reach $30.01 billion in 2026, growing at the Compound Annual Growth rate of 35.5% from 2019 to 2026.
The worldwide average rate of 12 billion kW every hour of every day of every year, more than 80% of the power generated, is being reprocessed or recycled through some form of power electronic systems. The growth of power electronics for electric vehicles market size is driven by several factors such as extensive demand for energy-efficient battery-powered devices and stringent emission regulation to reduce vehicle weight and emission.
A new initiative by the government in developing countries for balancing the pollution and emission of vehicles is also contributing to the market growth. However, the high cost of vehicles and complexity in designing and integrating advanced power electronic components in EV’s are predicted to impede the power electronics for the EV market growth. Nonetheless, rapid advancements and innovation in vehicle battery technology are anticipated to create profitable growth opportunities for the power electronics for the EV market.
The surge in demand for energy-efficient battery-powered devices, stringent emission regulations to reduce vehicle weight and emission, and goverm=nment initiatives to balance environmental pollution and vehicle emission are expected to drive the growth of the power electronics
For the electric vehicle industry. However, the high cost of vehicles and complexity in designing and integrating advanced power electronic components in electric vehicles hinder the power electronics for electric vehicle market growth.
Technological advancements in vehicle batteries and an increase in research activities are expected to create profitable growth opportunities for the power electronics for electric vehicles. Also, power electronics support high input impedance and proved parallel current sharing, which increases the adaptation of power electronic components in EV.
The global power electronics for electric vehicle market analysis is based on various applications including inverters, converters, and onboard chargers. Among these applications, the onboard charger is anticipated to witness the highest growth rate during the forecast period, owing to rapid innovations and developments in electric components used in an electric vehicle for better performance and energy efficiency.
Over the past few years, EVs have witnessed an increase in demand due to their lightweight nature and higher efficiency. This has led to a rise in the installation of e-axles and in-wheel hub motors in all types of EVs. Furthermore, such installation helps reduce CO2 emissions and the space required for engines. The traditional gas-powered vehicle makes use of an internal combustion engine to generate power.
1. In terms of revenue, the inverter segment impacted the maximum power electronics for electric vehicle market share in 2018 and is expected to maintain its lead throughout the forecast period.
2. The onboard charger segment is expected to grow at the highest CAGR during the forecast period.
 Glaser, J. (2017). How GaN Power Transistors Drive High-Performance Lidar: Generating ultrafast pulsed power with GaN FETs. IEEE Power Electronics Magazine, 4(1), 25–35. doi:10.1109/mpel.2016.2643099