Over the past few years, companies like Waymo, Tesla, Uber, and a host of startups have driven home the promise of self-driving cars. While consumers continue to wait for safe autonomous vehicles that can be summoned with a few taps on a smartphone, one area of autonomous vehicles has quietly moved into focus: autonomous trucking. Self-driving semi trucks and trailers are being developed by small startups in cooperation with shipping giants like UPS, and it is much more likely these vehicles will hit the road before their smaller counterparts.
Telematics systems have already seen increased use in 2020 as more companies rely on fleets of vehicles in order to comply with stricter health and safety requirements. Autonomous trucks will depend more on telematics systems as they perform many of the critical data acquisition tasks that help ensure safety, security, and efficiency within fleets of self-driving vehicles. Automotive OEMs will play a lead role in expanding the capabilities of telematics systems as more self-driving vehicles find their way onto the road.
How Autonomous Trucks Will Use Telematics
In addition to lidar, radar, cameras, and complex algorithms for controlling self-driving trucks, all autonomous vehicles need telematics systems to collect vehicle data and help make critical decisions in real time. Data from telematics systems will also be used off-vehicle for fleet tracking and critical decision-making. These systems are one of many technological advances that will help make autonomous trucking safer, more reliable, and more efficient.
Real-time Data Acquisition and Transmission
Telematics systems interface with the network of in-vehicle sensors and ECUs to gather and process data. Some of the data that can be collected include:
- Real-time location and route tracking by interfacing with the vehicle’s GPS system
- Vehicle speed and performance by monitoring engine data
- Vehicle behaviour over time (breaking, acceleration, reaction to obstacles, etc.)
- Diagnostic signals gathered from vehicle sensors and ECUs
These signals are already used in a variety of ways on and off the vehicle, particularly in driver performance monitoring programs and to develop highly accurate insurance policies for vehicle fleets. Driver performance monitoring allows fleet operators to spot inefficient driving behavior (e.g., excessive speed or idling) and dangerous driving behavior that would otherwise go unnoticed. Unauthorized vehicle use can also be spotted, such as cases of theft or deviation from a planned driving route.
As vehicles become more autonomous, important data like weather conditions, road conditions, location, speed, and vehicle health still need to be monitored to increase vehicle efficiency and ensure safety. All these data will be leveraged in increasingly complex AI models to aid greater automation, ultimately relieving stress on the vehicle operator on long-haul trips.
Beyond Level 1 Automation
Today’s newer trucks include ADAS capabilities that have converged with telematics systems. Data collected by telematics modules is already used in conjunction with ADAS systems to activate safety mechanisms in Level 1 autonomous vehicles. However, as automation in autonomous trucks increases and the driver can take a less active role in controlling the vehicle, the dependency on telematics will continue to increase.
The data collected by telematics systems is also useful for orchestrating actions among multiple trucks via VANETs. As a fleet of autonomous trucks travels, these vehicles will share their telematics information and orchestrate driving, ultimately without human intervention. In addition, autonomous trucks will become aware of actions taken by human-driven cars that transmit their telematics data, reducing the need to interpret sensor data to determine the actions of other vehicles. This will be enabled over short-range wireless links with on-device WLAN capabilities, creating a safer situation on the road for human-driven vehicles and groups of autonomous trucks alike.
Data gathered on the vehicle and from the surrounding environment can also be used off-vehicle for a host of management tasks. As more vehicles integrate cellular services into their hardware stack, data can be transmitted back to the cloud, allowing remote operators and the truck manufacturer to see important operating data in real time. Fusion of these data with existing traffic and infrastructure data can be used in load tracking and scheduling with the goal of ensuring most efficient routing and reducing vehicle dead time.
Predictive Analytics for Fleet Maintenance and Management
ICT tests identify short circuits, open circuits, poor or weak solders and incorrect, missing or misoriented The wealth of data gathered with telematics systems can be used in AI models for predictive analytics. A major area where AI can be used with telematics data is in preventative maintenance, where critical systems are monitored during the life of the vehicle and the chances of vehicle faults are determined. The goal here is to predict breakdown and required maintenance in order to keep a vehicle operating at peak efficiency. These models can also be leveraged to identify specific emergency situations, allowing the vehicle to provide much greater detail to first responders.
Implementing predictive analytics is aided by a wealth of new AI-centric SoCs that are just starting to reach the market. The newest class of integrated circuits have smaller footprints than the most powerful MPUs, GPUs, and FPGAs that are currently used for on-device AI. These components allow lightweight AI models to be performed on telematics systems without relying on the cloud, a move that helps ensure reliability and accuracy while reducing overall system costs.
Vehicle Testing and Qualification
As autonomous trucks become more connected and more automated, they will need to undergo rigorous testing in order to prove safety and operational efficiency. The real-time data acquisition capabilities of telematics systems allows test vehicles to be tracked and monitored during long-haul vehicle tests. The wealth of data gathered by these systems can then be used to evaluate current and future systems in the vehicle with the goal of ensuring safety and conformance to regulatory requirements.
As more vehicles become electrified, including autonomous trucks, battery and trip management will need to be approached in tandem. Power management in an electric vehicle requires constantly monitoring data from the battery management system with a telematics unit. Some of the key metrics that can be monitored include:
- Energy consumption over distance
- Battery state of charge
- How power consumption relates to driving behavior
- Battery diagnostics to ensure safety
The final point above is quite important as innovative battery companies continue to develop new chemistries to support electrified autonomous trucking fleets. By monitoring the power management and delivery systems in autonomous trucks, freight companies and vehicle manufacturers can use the data for important management tasks. Examples include optimal trip planning, load planning, and defining the vehicle’s driving style with the goal of balancing minimized energy consumption and maximized driving range. As part of a transition to electrification, fuel inefficient vehicles can be identified and removed from the fleet early based on their performance data.
Challenges in Telematics Integration
Although telematics systems will play a key role in driving adoption of autonomous trucking, there are still outstanding challenges to solve on the device and at the platform level:
- Large volumes of data: Integration of video telematics in commercial vehicles creates a huge volume of data, requiring unique processing, storage, and management solutions. Some of these will be addressed at the board and firmware level with denser memories and faster SoCs, but data management also requires optimized end-to-end communication.
- Lack of standardization: Current software solutions for automotive telematics are somewhat disjointed, and companies still need to build custom dashboards to monitor data gathered from multiple incompatible systems.
- VANET management: As more autonomous trucks hit the road and become connected into large fleets, vehicle orchestration becomes more of a challenge. This is partially a cloud management problem, as well as a network topology problem, both of which will require identifying and using only critical portions of data that can be collected from telematics systems.
OEMs that want to bring their telematics solutions to market should partner with an experienced electronics manufacturer that understands the autonomous trucking landscape. The right manufacturer can help companies navigate the complexities of volume production while also helping to ensure circuit boards for telematics products are reliable.