Like it or not, Autonomous Vehicles are coming – and, they’ll be here before you know it. That is unless cautious legislation and special interests decide to slow their deployment. In fact, many vehicles already on the road incorporate key elements necessary for Level 2 autonomy (also known as Advanced Driver Assistance Systems or ADAS). Some limited vehicles achieve Level 4 autonomy today, but still maintain steering wheels and pedal controls for many situations (such as the Tesla S Autopilot system). More Level 4 and some Level 5 (fully self-driving without human intervention) vehicles should arrive in the next 3 – 5 years, and perhaps sooner.
Level 5 vehicles will no longer require steering wheels or driver controls, as they will be fully capable of navigating streets and highways without a driver. Auto manufacturers and suppliers, as well as local, state, and federal governments, and nearly every business are trying to understand what this all means to them. It has been suggested, and I fully believe, that the introduction of Autonomous Vehicles (beginning roughly by 2020+/-) will initiate a societal paradigm shift as big as (if not bigger than) the introduction of motor vehicles to former dirt horse and buggy trails. Nearly everything, and everyone, will ultimately be impacted.
The potential benefits are many; improved safety, a more efficient and flexible transportation network, accessibility to mobility for those who could not drive, the potential to fully eliminate motor vehicle fatalities, injuries, and accidents, and more. But the reality is that this transition will take many decades before such benefits are realized. It may ultimately also cause many things to become obsolete and unnecessary. For example, some everyday items we take for granted that may disappear include things like auto insurance, the DMV, steering wheels, pedals, headlights, turn signals, traffic signs, traffic lights, taxi/Uber/Lyft drivers, traffic cops, driving schools, traffic court, tickets, speed radar, parking lots, parking meters, garages, marked traffic lanes, toll booths, toll takers, drunk drivers, and dare I say, Accident Reconstructionist and Forensic Experts?
Well, probably not the last two. In the decades it will take to fully eliminate humans behind the wheel, there will be some level of natural chaos associated with the blending of human drivers and varying levels of Autonomous Vehicles (just as there were with horses and horseless carriages). As this transition matures, experts will continue to be in demand for existing, emerging, and new technologies that are rapidly being deployed for autonomous vehicles. At present, there is considerable debate over whether we should move immediately to Level 5 (full) autonomy as soon as possible, or stage this evolution more gradually, with a capable driver still behind the wheel to take over in emergencies for some prescribed time period.
A recent row in San Francisco resulted in Uber shipping its allegedly autonomous Volvos to Arizona for further testing after San Francisco and the California DMV demanded that they apply for Autonomous Vehicle Testing permits. Uber’s position was that they were no more autonomous than a Tesla S with Autopilot engaged, as they had live drivers behind the wheel (at least one of whom ran a red light while driving and was captured on video). Oops. While technology is moving quickly it is possible that legislation or regulatory issues could slow its evolution, particularly if incidents involving autonomous vehicles rise.
Recent studies indicate that it will be difficult (if not impossible) for live drivers to safely take over driving when a vehicle’s autonomy fails or says, “your turn,” perhaps explaining the red light violation. In fact, many agree that the complexity of developing automation that can safely hand over control to an unpredictable live driver in emergency situations is more complicated than simply automating the autonomous driving response for the safest outcomes by default. A number of industry players including Ford, Waymo (Google), Mercedes, and a growing list of others are increasingly coming to this conclusion, and may jump directly to Level 5 for production vehicles.
Which brings me back to experts. As an expert in Automotive Electronic technology matters, intellectual property, and as a forensic expert in vehicle accidents resulting in death or injury, I can safely and assuredly predict that this will NOT go smoothly across the board. As dependence on advanced electronics, and to a greater extent, the software to manage all this has expanded in complexity, a perfect storm of design vulnerabilities, poorly executed and tested software, the inability to produce perfect products, and inadequate, ill-conceived and poorly written regulatory requirements is brewing. In the recent past, we’ve seen a slew of expensive and extensive recalls of major vehicle brands exceeding 7M vehicles impacted for software reasons alone. In addition, there have been several recent instances of alleged or admitted fraud and criminal behaviors by several manufacturers impacting another 11M vehicles involving embedded software.
Regardless, it will still take several years before cybersecurity measures are fully implemented that offer protection for on-board networks, connected vehicles, and consumers. Thus, the threat of bad actors hacking vehicles, as our government has confirmed, is quite real. I am personally involved in vehicle cybersecurity work at present, and find that many current production vehicles contain various undesired and hidden vulnerabilities due primarily to various software deficiencies that need to be corrected.
As I see the world of autonomous vehicles evolving, I see increasing need for experts in numerous expanding areas. Just a few of these include:
- Image Processing – For distinguishing objects around the vehicle at all times.
- Deep Learning/Neural Networks – Enabling vehicles to learn how to drive much as humans learn.
- Radar/Lidar – For sensing beyond the visual. Experts in this area will also be involved in managing how exponential increases in radar signals can be implemented effectively and safely.
- Software Development and Management – Managing multi-millions of lines of code to insure safety and security is a growing industry challenge.
- Human Use Study – People will change their behaviors during this transition to fully autonomous vehicles, and anticipating how and why will become increasingly relevant in developing effective and safe technologies and systems.
- Ethical Engineering – Programmers must help vehicles make ethical decisions in milliseconds in emergency situations. Understanding not only the technical issues, but also the ethical choices in programming, is akin to software development in defense applications.
- Networking & Connectivity – Onboard and wireless vehicle networks are rapidly evolving, and both are critical to the necessary infrastructure for autonomous vehicles. Methods to smoothly move in and out of connectivity, and move from server to client operation and back will be crucial to the success of autonomous vehicles.
- Land Use Policy – Most cities today devote roughly 30% of their land area to park vehicles. Autonomous vehicles can potentially reduce this requirement substantially for more efficient land use. Critical decisions about road infrastructure, parking issues, and societal impact will become increasingly relevant. This also spills over into Transportation Planning. Could autonomous vehicles also diminish or significantly impact the need for traditional public transportation?
- Cybersecurity – Connected and autonomous vehicles certainly elevate the need and requirements for adequate cybersecurity to insure their operation is not hacked or otherwise exploited by bad actors.