ADAS + V2X Trumps Self-Driving

The “talking cars” feature added to ADAS is expected to boost safety beyond current levels.

By Lou Frenzel, Contributing Editor
[Sponsored by PRQA]

Communications-enhanced ADAS will deliver better safety, but creates a software-development problem.

The primary objectives of the DOT’s Intelligent Transportation System (ITS) and the related National Highway Transportation Safety Administration (NHTSA) regulations are to improve vehicle safety and reduce accidents, injuries, and deaths. That is also the goal of self-driving vehicles. The hypothesis is that by taking the human out of the process, driver error will be minimized and safety will be significantly improved. Of course, that remains to be seen.

In the meantime, a very positive alternative is to further enhance current advanced driver assistance systems (ADASs). One of the forthcoming enhancements is the vehicle-to-vehicle (V2V) and vehicle-to-everything (V2X) communications systems that have been proposed and even mandated, but not yet implemented. While self-driving cars are in our future, ADAS + V2X may be a more acceptable option for many drivers. A key issue is developing the software to successfully integrate these subsystems. One emerging software solution is AUTOSAR, which will be discussed later in this article.

Advanced Driver Assistance Systems

ADASs have a growing presence in most new vehicles. Manufacturers have been adding these systems as the government demands greater safety for motor vehicles. The manufacturers are okay with this, as they regularly need new options and features to help sales. Buyers also are showing their interest and acceptance. Of course, as ADAS additions increase, prices will come down, making them an integral part of the vehicle cost. While consumers and the insurance companies do not show a total buy-in at this point, manufacturers are moving forward with new products and features. Even aftermarket products are being developed and sold. The electronics industry is booming as a result, as new cars are loaded with electronics that now dictate major design decisions.

ADAS features vary with the manufacturer but usually include a core of:

  • Backup camera
  • Lane-keeping steering
  • Blind-spot detection
  • Adaptive cruise control
  • Automatic braking

Other features already offered on some models and perhaps to come on others are:

  • Parking assistance
  • Adaptive front lighting
  • Drowsiness monitoring
  • Pedestrian detection
  • Traffic-sign detection
  • Alcohol detection
  • Night vision (IR)

As more functions are added, the number of embedded processors increases and software development continues to be the primary design effort. The ongoing challenge is to fully integrate the new functions into an already complex embedded control system.

Introduction to AUTOSAR Coding Guidelines
ADAS + V2X Trumps Self-Driving

AUTOSAR (AUTomotive Open System ARchitecture) aims to standardize and future-proof basic software elements, interfaces and bus systems, to help vehicle manufacturers manage growing system complexity while keeping costs down. This White Paper gives a brief overview of the new AUTOSAR Coding Guidelines and offers guidance on how to comply with them.

Download free white paper

How ADAS Works

Figure 1 shows a concept diagram for most ADASs. A cluster of sensors like video cameras, 77-GHz millimeter-wave radars, LIDAR, and ultrasonic transducers feed their digital outputs via multiple serial interfaces to processors that store the vision data and process it (see sidebar, “Lidar”). The vision processors prepare the data for fusion processing, which combines the inputs from the multiple sensors to produce feature recognition of nearby objects as well as their distance, motion, and status. Fusion processing then formats the data that will be handled by processors. They in turn make decisions and initiate the control of selected driving functions or provide driver notification by visual, audible, or haptic signals.

This concept of advanced driver assistance systems shows the sense-interpret-action process flow.

The whole objective of ADAS is to provide a continuous picture of the environment surrounding the vehicle. This vision of the area around the vehicle is what the driver typically experiences. But ADAS offers improvements that give even more information. With a reliable status report of the space around the vehicle, the driver and/or the ADAS can respond appropriately.

The Connected-Car Addition

One forthcoming new technology to be adopted is vehicle-to-vehicle (V2V) and vehicle-to-everything (V2X) radio communications systems. Such a system has been mandated by the Department of Transportation and its National Highway Transportation Safety Administration. This system allows cars to talk to one another and exchange data, such as location, direction of travel, speed, brake and accelerator status, pending turns, and other facts. This information can then be analyzed and used to avoid collision. The vehicle-to-infrastructure (V2I) subsystem will also let cars connect to road-side units (RSUs) that can supply additional data like signal light status, traffic congestion, road conditions, weather alerts, construction, and other conditions that can provide safety as well as convenience.

This communications system, called Dedicated Short Range Communications (DSRC), has already been fully defined and vetted for V2V and V2I. It operates in the unlicensed spectrum around 5.9 GHz and is a variant of Wi-Fi called 802.11p. Chips and modules are already available and extensive testing has been performed. This system will enhance the existing ADAS capabilities, further improving safety.

However, V2V and V2I have yet to be included in vehicles, despite the government mandate. The reason is that an alternative communications system has been proposed and touted as being superior. That system is based on the cellular phone system and is called C-V2X. It is designed to initially use the popular LTE 4G radio system and eventually evolve into the 5G system under development. This controversy has stalled adoption, as the government and industry attempt to agree on which format to embrace.

The V2V/V2X systems will connect to and interact with the ADAS, providing intelligence beyond the short-range environment covered by the ADAS sensors. This remote information will be melded with the situational awareness provided by the ADAS, providing even greater safety.

It’s a Software Problem with a Solution

ADAS development is a massive software project that includes special algorithms as well as artificial-intelligence (AI) techniques, such as machine vision and deep learning. It would be helpful to have a complete development platform that integrates the following: popular processors, buses and interfaces, common software in C and C++ programming, and AI hooks to speed and simplify development. That development platform does exist and has been making automotive software easier to develop for over 10 years. It is called AUTOSAR.

AUTOSAR, for AUTomotive Open System ARchitecture, is a standardized and open software architecture created jointly by auto manufacturers, OEM suppliers, and software tool developers. Its primary objective is to introduce a standardized layer between the specific application software and the processor(s) of the electronic control unit (ECU). This makes the software independent of the selected microcontroller or auto manufacturer, making it reusable in different ECUs.

There are three major trends to consider. The first is the approach of reducing the number of domain-specific processors and creating architecture around a smaller number of multi-core processors. Second, the major new and existing subsystems are increasingly being networked. The network of choice is Gigabit Ethernet, which provides the needed speed and bandwidth to make the combined systems function together. Finally, security is becoming a major consideration—especially if the auto systems are ultimately connected to the internet or other networks, where they may be exposed to hacking.

AUTOSAR is basically a collection of specifications. It includes software modules and interface definitions combined with a standard exchange format that defines a common development process. It comes in two forms: the Classic Platform and the Adaptive Platform. The Classic Platform is used in common vehicle functions including those that are safety related and require deterministic capabilities. The Adaptive Platform is a newer version, which addresses the additions of connected and autonomous vehicles. The target operating system is POSIX and other Linux-based OS.

The C and C++ languages are the most common for programming automotive systems. C++ is the suggested language for the newer Adaptive Platform. Software research firm PRQA recently joined the AUTOSAR development effort to establish guidelines for the use of the C++14 language in critical and safety-related systems. Specifically, it created the guidelines to be used as extension of the existing Motor Industry Software Reliability Association (MISRA) C++ standard used by most manufacturers.

The firm also released new versions of the QA-C and QA-C++ automated static code analysis tools. Automated static code analysis software removes coding defects before compilation, thereby cutting down on test and debug time and delivering projects to completion faster with improved error-free code. Such tools are almost mandatory today, as manual code reviews are not generally regarded as a viable development process because of the increasing complexity of automotive systems.

Other Design Issues

While software is the core development challenge, here are some additional design issues to consider as new ADASs are created:

  1. Expedite the incorporation of V2V, V2I, and V2X technologies. Adopt DSRC for V2V and the cellular option for V2I and other V2X uses.
  2. Address the human-machine-interface (HMI) problem. The HMI is critical to making ADAS fully successful. Notifications must be non-distracting. Bigger and brighter LCD dashboard screens may not be the answer. Consider heads-up displays (HUDs) and voice control/response, for example.
  3. Solve the testing problem. Some ADAS testing companies, like Konrad and Vector, have emerged to serve the industry. Simulation software is also becoming available to evaluate and test ADAS effectiveness before and after implementation. Multiple vendors already offer test instruments and software to fully vet the DSRC and C-V2X systems.
  4. Provide improved documentation and driver instruction in how to use ADASs.

We will eventually get self-driving vehicles. In the meantime, safety will improve as more and better ADASs become available. V2V and V2X technology will make ADAS an attractive alternative to full automation. At the same time, software like AUTOSAR is making it faster and easier to integrate new ADAS features.

LIDAR

Light detection and ranging or LIDAR sensors, based upon scanning lasers, are not currently used in ADAS because of their high cost. Their ability to generate a 3D image over a full 360-degree range plus improved weather performance over other sensors make them superior. Pricing is coming down as the technology is developed. Affordable LIDAR is critical to self-driving cars and will greatly enhance ADAS.

Introduction to AUTOSAR Coding Guidelines
ADAS + V2X Trumps Self-Driving

AUTOSAR (AUTomotive Open System ARchitecture) aims to standardize and future-proof basic software elements, interfaces and bus systems, to help vehicle manufacturers manage growing system complexity while keeping costs down. This White Paper gives a brief overview of the new AUTOSAR Coding Guidelines and offers guidance on how to comply with them.

Download free white paper

LINKEDIN
EMAIL
FACEBOOK
TWITTER

Start typing and press Enter to search