California State University, Long Beach
Inside CSULB Logo

Self-driving Vehicles Are Here

Published: November 1, 2016

Robots in motion with minds of their own–from drones above to cars below–will be the topic of a presentation by Computer Engineering and Computer Science’s Oscar Morales-Ponce when he speaks on cooperative mobile robots as part of the 45th Sonoma State University Computer Science Colloquium Series on Nov. 17.

Cooperative mobile robots are autonomous entities capable of self-coordinating their actions to solve common problems, explained Morales-Ponce, who joined CSULB in 2015.

“For example, a set of mobile robots can be used to patrol a protected area more efficiently than using only one robot,” he said. “Other examples are intelligent vehicles. In this scenario, vehicles can self-coordinate maneuvers such as crossing uncontrolled junctions with minimum time or safely changing lanes. This presentation will describe some challenges and solutions that arise with the use of cooperative mobile robots.”

Morales-Ponce is an expert on autonomous vehicles whose research focuses on the design and analysis of cooperative algorithms to self-coordinate the actions of distributed systems like self-driven cars. The key is coordination.

“Drones are like people,” he said. “If you have two people covering an area, they can cover more. It is the same with drones. The problem is how to coordinate their movements. Any search may lose time if the same place is searched twice. It is important to perform this coordination in a very efficient way.”

The key is “cooperative algorithms.” Because of their complexity, most modern systems are reliant on scheduling algorithms for efficient multitasking. Morales-Ponce’s research focuses on the design and analysis of cooperative algorithms to self-coordinate the actions of distributed systems like self-driven cars.

“Cooperative algorithms are distributed computing,” he explained. “The idea is that different computing systems must do what they are supposed to do and never collide. They tell each other when one moves in this direction and the other moves in that direction.”

The speed of communication between vehicles is vital.

“Even at slow speeds, autonomous cars exchange several messages every second,” he said. “Autonomous vehicles need to react immediately. With standard Wi-Fi, up to eight messages can be sent in a second altogether. If there are two cars communicating, they can send three to four messages a second. With homemade technology, there can be hundreds of short messages a second altogether. It is possible to control several drones with communication that fast.”

This high-speed decision-making is fundamental to self-driven cars changing lanes or intersecting.

“How self-driven cars intersect without smashing into each other is one of the main topics of my research,” he said. “Communications must be reliable. We must be sure one drone receives messages from other drones using either explicit or implicit acknowledgements.”

For instance, Morales-Ponce describes what could happen if several self-driving cars reached an intersection together.

“The idea is for the cars to inform each other their trajectory,” he said. “One car will tell the other, ‘This is my path and I will reach the crossing at this time.’ If the other car doesn’t respond, it doesn’t know what time the other car will reach the intersection. In that case, they could crash. If one car fails to elicit a response from another car, it knows that there is a problem and will react by approaching with caution.”

Self driving bots
PHOTO COURTESY OF OSCAR MORALES-PONCE

Other forms of transportation may be open to self-driving.

“Mining transportation or aircraft may benefit from computer consensus,” he said. “My Swedish research used the same algorithm for airplanes that ground transport does. The goal is to pack more planes into existing air space. Too many planes limit the amount of flight. The idea is to fly planes closer and, to do it, they use the same algorithm that ground vehicles use.”

Morales-Ponce rode a self-driving Volvo during his Swedish research and felt it was strange but interesting at the same time.

“There was the car making its own decision,” he said, “so that was a little bit strange.”

He was a Postdoctoral Scholar at Chalmers University of Technology from 2012-14 and at Carleton University from 2011-12. He obtained his Ph.D. in computer science from Canada’s Carleton University in 2011.

There are many challenges to self-driving transportation including the law.

“Who is responsible for a collision? Driverless cars are allowed on certain parts of Sweden’s roads. There will be places like that in the U.S. someday,” he said. “The insurance companies have taken a strong interest. Who is responsible for a crash? Is it the company that made the car or the driver?

“There are many levels of autonomous driving,” he added. “There are five levels from the highest autonomous levels without option for human intervention to the lowest advance cruise controls. Advance cruise control is a kind of autonomous control in that the car takes control of its speed.”

Morales-Ponce believes self-driving cars have a long future and sees them doing well in the big cities where people follow the traffic rules most of the time.

“But,” he noted. “I don’t see them in cities where people drive very aggressively. Italy, China or Argentina will not welcome autonomous cars.”