Fleet Sizing: Autonomous Cars Needed to Re-Image All Continental U.S. Roads in One Year

You are a product manager on Google Maps. The team wants to refresh Street View imagery across the entire continental United States using self-driving cars. How many cars would it take to re-map the whole area within one year? ```hint Two halves of one equation Turn the job into total hours of driving, then work out how many useful hours one car can deliver in a year. Daylight and weather limit the second number more than the car does. ``` ### Clarifying Questions - Does re-mapping mean driving every public road once, including unpaved rural roads, or only roads that already have street-level imagery? - Does capture need daylight and dry weather, or can some sensors keep working at night? - Can the cars drive autonomously on every road type and in every region, or will some areas need a human driver? - Can cars move between regions during the year, for example heading south while northern roads are under snow? ### What a Strong Answer Covers - A precise definition of the job (which roads, how many passes) stated as explicit assumptions - A fleet equation that converts road miles into driving hours and divides by productive hours per car - A defensible estimate of U.S. road mileage and average capture speed, with a sanity check - Realistic limits on each car's year: daylight, weather, charging, maintenance and a spare margin - A range, plus the assumptions that move the answer most ### Follow-up Questions - If the deadline shrank to six months, or only cities needed a yearly refresh while rural roads were refreshed every few years, how would the fleet change? - How would you decide whether a full annual re-map is worth its cost in the first place? - Which parts of the plan change if the cars cannot run autonomously on unpaved or mountain roads?