When Should the Chicken Cross the Road? - Game Theory for Autonomous Vehicle - Human Interactions

Autonomous vehicle control is well understood for local- [15], good approximations exist such as particle �ltering, ization, mapping and planning in un-reactive environ- which make use of large compute power to draw samples ments, but the human factors of complex interactions near solutions. stood [16], and despite its exact solution being NP-hard with other road users are not yet developed. Route planning in non-interactive envi- ronments also has well known tractable solutions such as This po- the A-star algorithm. Given a route, localizing and con- sition paper presents an initial model for negotiation be- trol to follow that route then becomes a similar task to tween an autonomous vehicle and another vehicle at an that performed by the 1959 General Motors Firebird-III unsigned intersections or (equivalently) with a pedestrian self-driving car [1], which used electromagnetic sensing at an unsigned road-crossing (jaywalking), using discrete to follow a wire built into the road. Such path follow- sequential game theory. The model is intended as a ba- ing, using wires or SLAM, can then be augmented with sic framework for more realistic and data-driven future simple safety logic to stop the vehicle if any obstacle is extensions. The model shows that when only vehicle po- in its way, as detected by any range sensor. sition is used to signal intent, the optimal behaviors for open source systems for this level of `self-driving' are now both agents must include a non-zero probability of al- widely available [6]. lowing a collision to occur. In contrast, This suggests extensions to problems that these vehicles will face around interacting with other road users are much harder reduce this probability in future, such as other forms of both to formulate and solve. Autonomous vehicles do not signaling and control. Unlike most Game Theory appli- just have to deal with inanimate objects, sensors, and cations in Economics, active vehicle control requires real- maps. time selection from multiple equilibria with no history, They have to deal with other agents, currently human drivers and pedestrians and eventually other au- and we present and argue for a novel solution concept, meta-strategy convergence , suited to this task.