Autonomy, Safety, Electric Transportation – Related? Complementary? Conflicting?
By Philip Krein, Chair, IEEE Transportation Electrification Community
Sometimes it seems as though the problems associated with autonomous vehicles are not effectively posed. Much of the literature seems to emphasize questions along the lines of
“How can an autonomous vehicle successfully negotiate traffic and obstacles to move from point A to point B?”
This, in essence, is a robotics challenge, and reflects impressive advances in robotics over the last few years. It sounds good, but consider a different statement:
“How can a vehicle be made to function substantially better and more safely than with any human driver, over the full range of its intended application and emergency situations?”
Today, it seems as though much of the research emphasizes the first statement, adding efforts to solve the problem less expensively and with simplified approaches. In some sense this seems to “put the cart before the horse.” The general public seems to like autonomous vehicles, but safety concerns are often expressed, along with ethical dilemmas and many other considerations. These, perhaps, could be addressed by considering the second statement. Obviously both can be taken together, but they seem rather orthogonal.
What can be done, and in particular, to what extent does transportation electrification have impact on either statement?
- Engineers will need to prove out driving performance across the board better than with the most expert human driver. The public might not require perfection, but certainly a “better than any human” target can get attention. This objective must not limit sensing capability, computing performance, actuation, or any other attributes needed to achieve success.
- Next, determine what features contribute most effectively to enhanced driving performance and determine how to introduce them most effectively into vehicles.
- Finally, add pathfinding ability, with enough intelligence to avoid problems such as map data conflicts and the like, for the “point A to point B” objective.
You might notice that the objectives raise a broader question: “What tools are most effective for improving driving performance?” These could include a human driver in the loop, or not, as the technology requires. There is substantial emerging work on “driver augmentation,” from heads-up displays and infrared night imaging to active cruise control, lane tracking, and automatic braking. Sometimes these seem a bit ad hoc, but certainly they are known to make driving safer.
What about electrification? Electric drive trains and electric actuators can be used to enhance dynamic performance, provide rapid torque response and active wheel control, make use of steering actuation and active suspensions, prevent some unsafe operating modes, and many others. A simple example is a drive-by-wire pedal system built in the mid-1990s in which conflicting pedal signals were interpreted based on a “brake dominant” strategy. In such a car, a stuck accelerator or equivalent sensor problem can be overcome by pressing the brake. Other researchers have used skid steering dynamic controls to enforce rapid directional control with multi-wheel electric actuators.
Does the combination of electric drive and robotics help in moving from point A to point B? Yes, since this can be done with a reduced environmental footprint and simplified vehicle systems. The links between electric drives and robotics challenges seem second order, however.
Can the combination of electric drive trains and “make a vehicle function better than with a human driver” improve global challenges of personal mobility and transport? Many of us believe the answer is an emphatic YES. Rapid actuation, enhanced dynamic performance, better control of power, and other fundamental advantages of electrification shine here. The IEEE Transportation Electrification Community is dedicated to a broad view of transportation challenges.