Paul AriasAutonomous truck: radar and GPS on the roof and cameras on the frontPaul Arias
Driverless cars do not get tired, do not get distracted and do not get drunk. They promise to transform urban mobility. Such cars are able to move without a driver behind the wheel, under the command of a computer control system connected to sensors and other equipment. Guided by the instructions of the user, they go from one location to another. Along the way, they must be able to obtain information about the environment, such as the presence of signs, pedestrians and other vehicles, as well as be guided by satellite systems. The technology for this purpose is under development through various projects worldwide, at universities and research centers and the automotive industry itself. Brazil is home to a few such projects at scientific institutions, at least one of which is in partnership with a vehicle manufacturer.
Scania has invested R$1.2 million in a project to develop a robotic truck in partnership with researchers at the University of São Paulo (USP), São Carlos campus. The model provided by the Swedish company to researchers for the work, begun in 2013, is the 9-ton G360. To make it autonomous, the team at the Institute of Mathematical Sciences and Computation (ICMC-USP), led by Denis Fernando Wolf, had to make mechanical changes and install sensors and electronic systems. “We coupled small electric motors to the wheel and brakes and placed a chip in the accelerator to control maneuvers and speed,” he says. The truck was also equipped with a stereo camera system and high-precision Global Positioning System (GPS), which gives the direction and position of the vehicle within a 5 centimeter (cm) margin of error. There are three cameras placed in front, each 1.5 meters (m) in height and positioned 20 cm from one another. Two of them always work in tandem, similar to the human eye, making it possible to estimate the distance of objects from the image.
There is also radar to identify obstacles under low visibility conditions. According to Wolf, the two types of equipment complement each other. The radar detects the presence or absence of obstacles at long distances. The cameras are more accurate and sensitive and can identify horizontal traffic signs or signals in the traffic lane and distinguish colors and textures, as well as people and animals in the road. But these devices alone are unable to drive the truck, which requires a “brain,” which is a computer. Installed behind the driver’s seat, it receives the data from the sensors and processes it, which controls the vehicle in real time.
GoogleCar developed by Google: was built to be autonomousGoogle
Sustainable vehicles
Rogério Rezende, director of Corporate and Governmental Affairs for Scania Latin America, says that the partnership’s main objective is not commercial. “We invest in research and development projects around the world and the autonomous vehicle project, in partnership with USP São Carlos, fits into this mold,” he says. “The idea is to generate knowledge, which is critical to advancing sustainable transport.”
In addition to the truck, Wolf’s team is working on another project, the Robotic Smart Car for Autonomous Navigation (Carina), which began in 2011, with support from FAPESP and the National Council for Scientific and Technological Development (CNPq), within the framework of the National Institute of Science and Technology for Critical Embedded Systems (INCT-SEC). While the goal of the truck project was to demonstrate the potential to develop the technology in Brazil at the lowest possible cost, while preserving its functionality, Wolf says that Carina is a research platform. “The more sensors there are, and the more sophisticated they are, the better the conditions for doing cutting-edge research.”
Carina is equipped with two cameras (in stereo), precision GPS and an inertial unit (a kind of 3D compass). It also has a rotating laser transmitter on the roof, which emits 32 beams to scan and identify everything in the surrounding area, which creates a kind of 3D road map. In October 2013 the vehicle was tested on the streets of São Carlos, with no one in the driver’s seat (see FAPESP Issue nº 213). “It was the first in Latin America to be tested on urban roads,” says Wolf with pride. “Previously, it had been tested in only a very few cities in the United States, France and Germany.”
The distance that Carina should have covered was 3 kilometers (km), but the total traveled came close to 30 km. The next car trip is expected to take place by the end of the year. The car is being prepared for use as an autonomous means of transportation on the University of São Carlos campus, for those who want to test the technology that’s been developed. To do this, an application will be developed whereby those interested in the service can call the vehicle. Once in the vehicle, passengers will indicate the destination via a touch screen. After the trip, the car will return to its parking spot and wait for another call.
Researchers at the Federal University of Espírito Santo (Ufes) are also developing a robotic vehicle, the IARA. The project began in 2009 at the High Performance Computing Laboratory (LCDA), with a small robo-car and a broad objective: to understand how humans can drive using their visual ability and skill to control the car. This stage received funding from the Espírito Santo Research Foundation (FAPES). In September 2012 the team imported a hybrid car, electricity and gasoline, from the United States. This second phase, which will last until December 2015, is supported by CNPq. Parts to adapt the car for research, such as the steering wheel, brake and special gearshift for autonomous vehicles, arrived in Brazil ready-to-go. “We had to add the other equipment,” says Alberto Ferreira de Souza, the project coordinator. This includes a laser transmitter, similar to Carina’s, for the roof of the IARA. In addition, it has a precision GPS and is surrounded by multiple cameras grouped in pairs, which act like the human eye. The Ufes group developed the software that controls the car, which has already completed a circuit around the Ufes campus, a route of 3.8 km. The next goal for the vehicle is an autonomous trip from Vitória to Guarapari, 50 km away.
Léo RamosCarina at USP in São Carlos: rotating laser on the roofLéo Ramos
History and safety
A group of graduate students in the Department of Transport Engineering of the Polytechnic School of the University of São Paulo (Poli-USP) are working on projects involving autonomous navigation using small carts 80 cm in length. Professor Edvaldo Simões da Fonseca Júnior coordinates the group. “One research project uses GPS to move the vehicle between two points,” he says. “Another involved building a car that can travel within a confined space, such as a building where GPS does not work, guided remotely through a Wi-Fi network. It could be useful inside mines, for example.”
Fonseca has been active in this area for many years and remembers that robotic cars have a very long history. In the article “Autonomous vehicles: concepts, history and state of the art,” presented in 2013 at the annual conference of the National Association of Research and Education in Transportation (ANPET), co-authored with his colleagues Rodrigo de Sousa Pissardini and Daniel Chin Min Wei, Fonseca says that automated transportation was first discussed at the 1939 World’s Fair in New York City. It was anticipated that within 20 years the world be exhibiting “an automated highway system prototype, where the roads would correct human driving errors and prevent maneuvers that could be dangerous or impossible to complete.” Since then, research has evolved, particularly with the emergence of mobile robotics. According to the article, the Mercedes-Benz VaMoRs prototype, a van equipped with cameras and other sensors in which the steering and other components were controlled by computer commands, appeared in 1985. The vehicle could, autonomously, reach a speed of up to 100 km/h on roads without traffic. Since then, several other companies in the automotive industry, such as Nissan, Volvo, Volkswagen and BMW, have begun to develop robotic cars. One of the most advanced has been built by a newcomer to the industry, the technology giant Google. The company began its research in 2010 by adapting models already on the market. Today Google has its own car, which resembles the Fiat 500, but is smaller. Its vehicle also combines sensors with radar, cameras and GPS. Google’s cars have a device called a laser rangefinder (an optical device used to measure the distance between the observer and any point) installed on the roof, which creates a three-dimensional map of the environment.
Developing the technology and building driverless cars are not enough to make their use a reality. Before putting them on the road, there needs to be some discussion about the rules that will govern them as they travel from place to place not to mention a determination of legal liabilities in case of accidents. Although theoretically they are less prone to errors, there are no absolute guarantees. “The driverless car will be made to follow the rules of the road and not to cause accidents, but it could be involved in a crash and hit another vehicle or run over someone. Then who will be responsible? The owner or the manufacturer? Can the car be insured? We do not know yet,” says Fonseca.
Wolf raises other issues that need to be addressed. “The technology to build the cars is almost ready, but its large-scale adoption will primarily depend on the market,” he says. “We need to wait and see if the automotive industry is interested.”
Although they are not yet on sale—which should take a few years—robotic vehicles have contributed to subtly introducing automation to the chaotic traffic of large cities. Among the most striking examples, Fonseca cites cruise control, which maintains the vehicle at a selected constant speed, the ABS brake, which prevents wheel lock and skidding, automatic parking, which parks the car using sensors, and collision avoidance systems with radar and cameras to detect an imminent collision. Autonomy is coming little by little.
Projects
1. National Institute of Science and Technology for Critical Embedded Systems (INCT-SEC) (nº 2008/57870-9); Grant mechanism Thematic Project-INCT; Principal investigator José Carlos Maldonado (USP); Investment R$2,639,677.06 (for all INCT-SEC) (FAPESP/CNPq).
2. Collaborative Effort For Safer And More Efficient Transportation With Intelligent Vehicles (FAPESP-OSU/2013) (nº 2013/50332-0); Grant mechanism Regular Research Grant; Principal investigator Denis Wolf (USP); Investment R$21,660.00.
3. Carina Project – Intelligent Robotic Car for Autonomous Navigation (nº 2011/10660-2); Grant mechanism Regular Research Grant; Principal investigator Denis Wolf (USP); Investment R$55,753.20.
4. Project Carina – Location and Control (nº 2013/24542-7); Grant mechanism Regular Research Grant; Principal investigator Denis Wolf (USP); Investment R$61,412.95 and $15,840.10.
