Often times, curiosity drives people to inquire about my robotics research. Mostly, these are people who do not necessarily work in the field of robotics. These include my friends and family (mostly my mom!), my colleagues from different departments, my dentist, my dry cleaner, the cook at one of the lunch trucks that I frequent, and elementary school kids. Of course, I love explaining to them what I do, sometimes at the risk of boring them! While I have several areas of robotics research, one of my favorites is driverless robotic vehicles. 

The robotics community has always had a very progressive view of the future with an intention to create technology that benefits humanity. A few of the best minds in the world work on really hard problems that arise in this field. Somewhere among these rockstars, I humbly play my part in the pursuit of turning science fiction into reality. 

The fruits of the collective labor of this community are extraordinary. For example, in 2011-12, Google demonstrated autonomous ground-based mobility in an urban environment. By August 2012, the Google Driverless Car had logged more than 300,000 miles in the state of California. Sebastian Thrun, the lead developer of the car described the broader impacts of this technology as follows, ``We can reduce traffic accidents by 90%. We can reduce wasted commute time and energy by 90%. We can reduce the number of cars by 90%".  Recently, Newman and others at Oxford University have demonstrated a driverless car with similar capabilities. Several states in the United States have already passed legislation that allow driving licenses being issued to driverless cars. 

During the same period, Kumar and others demonstrated multiple quadrotors operating in an indoor workspace and maintaining formations that translate and rotate with time. This is an example where multiple autonomous vehicles coordinate to achieve a collective task. 

The video shown here is an example of 5 robotic cars navigating along their paths in a cluttered environment while remaining fairly close to each other. These tiny cars use algorithms that I developed as a part of my research for their safe navigation. The most exciting part about my work is that we are not too far from the day when these algorithms will transition from the laboratory environment to commercial driverless vehicles operating in the real world. 

These and many other developments are harbingers of a future where multiple autonomous vehicles will become an all pervasive concept with applications that improve our standard of living and lifestyles. Such vehicles will lead to a reduction in the number of accidents and commute times and improved fuel efficiency. These, in turn, will drive down the cost of public transportation, logistics, and supply chain management, thereby allowing transportation of personnel and goods to previously unreachable locations in record times. 

While exciting, such autonomous vehicle-based applications come with their own challenges. For example, these vehicles may need to plan their motions while in transit and in real-time. This requires their decision-making mechanisms to be dynamic and efficient. There are kinematic challenges as well, e.g. cars cannot slide sideways to parallel park and a fixed-wing aircraft has a nonzero radius. Design limitations such as finite battery and fuel capacity add to the complexity of the situation.

Furthermore, if multiple vehicles are operating in a common workspace (e.g., urban environments, highways, airways), some shared knowledge of their behavior is important. To obtain this knowledge, either one needs to have sensors or the vehicles need to communicate with each other or with central entities that can update other vehicles with the latest information about their intentions as and when needed.

The biggest challenges, however, will be the commercial viability of such vehicles and adoption by the general public. Specifically, we need to address issues of energy efficiency and safety. Efficiency requirements dictate that these vehicles use as little fuel as possible, to travel as far as possible, as quickly as possible.  Safety requirements dictate that these vehicles be capable of avoiding obstacles and not collide with each other. 

Despite all these challenges, the current propensity of innovation in this space points to a future that is indeed bright...and driverless!