Mapping Decisions: The Promise of Mapless-ness

This chapter considers the role of mapping in the development of autonomous vehicles, offering a comparison between two firms—Uber ATG and Waabi—taking different paths to establishing the cartographic coordinates of their respective enterprises. Considered a fundament for such work, autonomous vehicle developers must concern themselves with building ‘operational design domains’ (ODDs), mapping the full geographic extent of their autonomous vehicle operations, down to the final millimetre. Far from a purely scientific endeavour, defined only by safety protocols or technical requirements, the setting of ODDs—most certainly in the case of Uber ATG—are informed by economic decisions concerning viability, scalability, and profitably. This fastidious ODD work—manual route driving, data tagging, domain building—crystallizes the search for viable future autonomous ride-hailing markets. Other operators, however, offer a different cartographic approach, less concerned with the mapping of geometric phenomena. Waabi’s stated ‘mapless’ approach to autonomous driving, although far from a-cartographic, demonstrates a sensor data-driven approach seeking to eradicate pesky errors that enter into the developmental pipeline at the mapping stage. These attempts to offer innovative solutions to traditional engineering problems—such as the modular, ‘software stack’ approach—are typical of the competitive nature of autonomous driving.