The Driverless City Project
Abstract
Ubiquitous co-robots offer society a unique opportunity to reshape our transportation infrastructure. To do so, we must balance the needs of safety, usability, and aesthetics. This research will investigate these tradeoffs in an effort to understand how to turn twentieth century transportation infrastructure (e.g. highways, intersections, roads, sidewalks) into twenty-first century human infrastructure that addresses the needs of the entire community. There is a clear link between a robot’s ability to localize within an environment and that environment’s physical structure. Tall buildings can severely compromise GNSS1 signals; landmarks that are spaced too far apart decrease localization accuracy; and the same landmarks spaced too close together can introduce a higher probability of faulty measurement associations. There is also a clear link between a robot’s ability to localize itself and safety—robots that cannot guarantee that their pose lies within a given covariance envelope are safety risks to nearby humans and objects. This safety risk may be small enough such that one or two mobile robots operating among humans would rarely cause harm. However, as we approach the concept of “ubiquitous co-robots,” the probability of unsafe operation increases by orders of magnitude that, if left unaddressed, could be life-threatening and erode the public’s trust in co-robot technology.
To mitigate safety risk, one could shape the environment to maximize a robot’s localization safety, similar to the careful placement of radio-navigation equipment at airports for approaching aircraft. This could be done by creating ordinances that dictate the appearance of the streetscape so that self-driving cars, drones, and other mobile co-robots can guarantee their pose covariance envelope in the presence of faults as they navigate. However, modifying the environment to maximize co-robot safety could have negative and wide-ranging societal impacts if the process does not consider the needs of other stakeholders, such as pedestrians, cyclists, drivers, vehicle passengers, homeowners, business owners, robot manufacturers, and the government.
In response, a highly interdisciplinary research project led by Professor Matthew Spenko (Director, IIT Robotics Lab), Boris Pervan (Director, IIT Navigation Lab), and Ron Henderson (Director, IIT Landscape Architecture + Urbanism Program) have received National Science Foundation funding to study the relationship between landscape architecture, city planning, and mobile co-robot navigation safety and the impact that ubiquitous robots will have on shaping urban design. The findings of this research, currently entering its third year, will be a framework that will inform city planners, architects, and landscape architects how to plan and design cities in which ubiquitous robots safely interact with humans and also educate roboticists on how the technology they are developing impacts other aspects of society.
