summary

Google Car

Self-driving cars are normally found in fictional movies, but Google is about to turn fiction into reality with the development of a full-fledged self-driven car. This means that the car can steer, accelerate, and can stop by itself. Google’s software, known as the Google chauffer, has components that include mission planning, behavior, perception, and motion planning and vehicle control.

Design

The vehicle employs the use of artificial intelligence software that exhibits human intelligence that exhibits human behavior. It includes voice recognition, face recognition, natural language processing, game intelligence, artificial creativity, expert systems, among others.

The mission planner component determines the waypoint segments that the vehicle should travel so as to complete a mission. It uses information such as road networks, terrain profiles, and information gathered during missions. After the information has been processed, it outputs waypoints to the behavior module.

Perception is determined by algorithms that perform localization, object classification, and road detection. Sensors such as Lidar and Radar integrate information so that it can be used by planning and reactive components.

The behavior component enables the vehicle to follow rules. The rules may be intersection progression for ground vehicles or docking for surface vehicles. In the event of the rules conflicting, the Anti Selection Mechanism (ASM) evaluates the most appropriate behavior given the situation. The behavior integrator ensures that there is a winner for each rule, so a full profile for behavior can be generated at any time. This structure encourages greater modularity and specialization for particular applications.

The motion planning is the decision-making component between the vehicle controller and the behavior. It converts target points into a series of vehicle commands. The motion planning comes up with a navigation strategy so that set goal points can be achieved safely. It is further subdivided into lanes and zones. The lane navigation requires maintenance of strict boundaries and conforms to the motions of other vehicles. The zone navigation requires a balance of speed and steering according to object avoidance.

The vehicle control is responsible for maintaining closed loop control of the vehicle’s actuators, monitoring safety systems, and reporting system health. Mechanical control systems comprise of servo motors and relays, brake control, driving wheel control, and throttle control.

The concept and function of Google car is assisted by Google map, hardware sensors and artificial intelligence software. The Google map interacts with GPS and acts like a database of speed limits, upcoming intersections, nearby collisions, traffic report and gives directions. Google car contains hardware sensors that gives real time environmental properties, creates a fully observable environment. The sensors employed are called Mobileye N.V. They offer a wide range of driver safety solutions.

Setup

The GX3200 is Google’s third car model, an effort to produce a fully electric, fully autonomous vehicle. The car can accommodate four passengers and has room for three suitcases in the storage compartment. Each car acts as its own wireless base station, so internet connectivity can be accessed through Google’s WirelessGig service. The model has low weight and can travel for 750 miles on a single charge. It can stay for about 48 hours on standby mode. It can also dock in the nearest Google PowerUP station when not in use.

The sensors used are Light Detection and ranging (Lidar), Video Camera, Position Estimator, and Distance Sensor. The Lidar is a Velodyne 64 beam laser that rotates on the roof. It scans more than 200 ft in all directions so that it can generate a precise three-dimensional map showing the car’s surroundings. It is the heart of the system and scans up to a distance of 60 meters. The Lidar has optimal remote-sensing technology that can read lane markers and stop lines with an accuracy of 2 inches. The car takes the generated maps and combines them with high-resolution world maps to produce different data models that allow it to drive itself.

The position estimator is mounted on the left rear wheel of the car. It measures small movements made by the car and helps to accurately locate its position on the map. Through the use of GPS, it determines the location of the vehicle and keeps track of its movement.

The video camera is placed near the rear-view mirror. It detects traffic lights and helps the car’s onboard computers recognize moving obstacles like pedestrians, cyclists, and other cars. There are two cameras, one with a 45 degrees point of view and the other a 90 degrees point of view. It reads lane markings with a higher accuracy than the Lidar. It also detects upcoming traffic lights.

The Light Detection and Ranging (LIDAR) sensor measures distance by illuminating a target with light and using pulses from a laser. It uses ultra violet, near infrared or visible light to image objects. Physical features are mapped by a narrow laser. A 3D contour of the car’s immediate surroundings are continuously generated , helping the car to sensor its surroundings.

A position sensor provides altitude, latitude, and longitude, and the corresponding standard deviation. When the geostationary satellites providing the GPS correction are visible from the car, the unit enters the high precision GPS mode. Standard precision GPS mode occurs when no correction signal is available.

A Radio Detection and Ranging (RADAR) is a system that detects objects by using radio waves to determine the range, direction, altitude, or speed of objects. It can detect other motor vehicles, the weather formation, aircrafts, ships, spacecrafts, and terrain. The antenna transmits radio waves pulses which bounce off any object on their path, returning a part of the wave’s energy to the antenna which is normally located at the same site with the antenna.

The distance sensors are in form of radars, three in front of the vehicle and one in the rear. They help establish the positions of distant objects. It can determine distance from a short range and long range radar up to a maximum of 150 yards. They measure distance to various obstacles and allow the system to reduce the speed of the car or increase it depending on the situation.

The car’s operation is controlled by a computer whereby the desired effects re delivered by the use of electronic throttle control.

Capabilities

Semi-autonomous waypoint following includes lane keeping, automatic exit, automatic lane change, overtaking slow or stopped vehicles, automatic deceleration behind congestion on freeways, and automatic stopping at red lights.

Follow the leader is a concept that would be useful for the military and industries. It entails the first vehicle being driven by a human and the following vehicles would only need to follow it.

Area clearance is all about the car establishing free parking space. The automated car is able to identify available parking spaces and determine the best slot to park the vehicle.

Tele-operated (remote) control- high level decisions are provided by the operator on where the vehicle should go. In many situations, a high-bandwidth, low-latency communication link is either unavailable or technically impossible to provide. In order to avoid delays in the transmission system, the operator directs the vehicle based on a single screen image. The vehicle follows a prescribed path for some distance as a new image is being created.

Google car obeys rules of the road. The perception function of the vehicle assists it to perceive information such as traffic lights and process it in order to perform the required action.

Intersection behaviors (precedence, queuing) are determined by the behavior component of the vehicle. The vehicle, when faced with a number of decisions to make, has to determine the one that takes precedence over the others.

It interchanges lanes in order to avoid traffic. It does this through the assistance of the motion planning component of the vehicle. The vehicle is able to able to move from one lane to the other and back if need be.

In terms of oncoming traffic, the vehicle is able to expect oncoming traffic due to the motion planning component of the software. In case of traffic, the vehicle automatically reduces speed and subsequently increases speed when the road is clear and there is no obstacle in front.

It detects obstacles at the back and in front of a car while manually or automatically parking through the use of an ultrasound sensor. It reverses flawlessly and occupies an empty parking slot.

Vehicular communication systems – Vehicles may obtain information from other vehicles in the vicinity especially if it relates to traffic congestion and safety hazards. Vehicles and roadside units are used as communication nodes where they provide each other with information.

Automatic rerouting - this is done by updating the vehicle’s map through the use of sensory input. The lane-keeping system uses a camera in the rearview mirror to monitor lane markings. It gently moves the steering wheel if the car drifts too far to the right or left.

Keeping track of the vehicle’s location is enabled by the wheel sensors. Stability and anti-lock and braking systems are also assisted by the wheel sensors. The vehicle applies smooth brakes when it sensors an object in front of it.

Retro-traverse, a return to point of origin along the previously traveled path can be employed by the Google car. A pre-planned route of navigation can be set whereby the car may transport passengers to all the pre-determined locations and come back via the same route.

Sensor data obtained from the road surface as well as from raised buildings are obtained from a laser scanner and are then accumulated in an occupancy grid. The information is extracted from a stationary environment and is used to detect and estimate road surfaces.

Excellent road following by the Google car is one of its main features. It maintains a safer following distance from the car in front, accelerates, and comes to a halt more smoothly.

Detecting and classifying obstacles (people, vehicles, and others) through the assistance of the radar is a capability of an automated car. This in turn avoids accidents such as hitting pedestrians. The vehicle also avoids collisions with other vehicles and ensures that it does not hit buildings or other stationary landmarks.

Detecting and interpreting regulatory traffic signs are done with the use of the video camera. An automated car detects the traffic lights colour through the use of a video camera that interprets lights. The wide-angle camera registers pixels that are used by the computer to analyze and establish colour.