An autonomous vehicle is any kind of vehicle that does not require real-time instruction from a human operator to function. Making a vehicle autonomous involves giving it a significant level of internal decision-making capability. Autonomous submarines and cars have been developed for a variety of purposes.
For an autonomous vehicle to make useful real-time decisions, sensors are required to obtain feedback information from the environment. Some form of onboard computer, often a small microprocessor, can take this data and determine the vehicle’s state. This may be as simple as tracking the vehicle's current location, but also may involve other types of data measurement. The onboard computer then implements pre-programmed responses based on its current state to make the autonomous vehicle behave in intelligent ways.
Research into submarines with autonomous capability began some 50 years ago. These autonomous underwater vehicles (AUVs) are used by the oil and gas industries to map the seafloor before subsea infrastructure such as oil and gas fields are built. In the military, AUVs are used to help detect other submarines and monitor protected areas for new objects such as mines. Scientists studying lakes and oceans may find an autonomous vehicle useful for gathering visual and compositional data.
The Mars Exploration Rovers can be described as semi-autonomous because they can make a number of independent navigation decisions. This capability was designed because of the delay in round-trip communication times between the Earth and Mars. These delays can range from about 6 minutes to a full 40 minutes depending on the position of the planets. To help the rovers progress further in a given day, engineers programmed them to take general instructions from human operators. The rovers can also make some specific decisions on their own. Varying levels of “courage” and “cowardice” can be programmed into the rovers depending on how dangerous the local terrain is.
Recently, the United States Defense Advanced Research Projects Agency (DARPA) sponsored a series of competitions to develop a driverless car. The DARPA Grand Challenge featured teams competing against each other to get their car through an off-road course in the fastest time. In 2007, DARPA sponsored an urban challenge, which involved a course in an urban setting. It is commonly believed that the DARPA challenges were aimed at spurring the development of an autonomous vehicle capable of operating in hostile urban war zones. The application of a driverless car, however, has been suggested for civilian use as well.