Project Overview

Problem Statement and Motivation

It has been proposed that an automated highway network for robots should be created. The goal of this project was to build a scaled prototype of a taxi that could transport passengers around four stations on an automated highway for robots. Money was earned by the delivery of passengers but was balanced against the cost of the design of the robot as well as any fines for traffic or rule violation. The motivation was to design the robot that could make the most money at the end of a trial run.

Design Constraints

The robot had to fit the dimensions of 1’ by 1’. The robot had to be primarily built from the NXT kit. Each part of the kit had its own cost, which was subtracted from the final profit.

Figure 1: The Official Highway

The arena consisted of a 5x5 square with diagonals marked out on the floor using black tape. At the center of the arena there was an unmarked red patterned area. When entering this area robots were given a restricted speed of 165 mm/second. All of the robots had to maintain a minimum separation of 6”. At any given time the robot had to be over a marked section of roadway. The robot could not stop moving for more than 30 seconds, otherwise it is considered illegal parking and a fine would be issued. If the robot stopped moving for over 45 seconds, it would be disqualified. Throughout the competition the robot could only carry up to 3 passengers at a time. The robot was required to display the number of passengers being carried by their identifying color. In the event of a collision, the robot at fault would be disqualified. 

Pre-existing solutions

While there are no current systems exactly like this one, there are several systems addressing autonomous transportation such as Personal Rapid Transit systems and the Google Car.

Figure 2: PRT System currently operating at West Virginia University

Starting in 1970, multiple Personal Rapid Transit systems (PRT) were developed under a government funded initiative to build cheaper and more efficient public transportation. PRT's consist of mainly off the shelf parts and systems, such as chassis from existing commercial vehicles, making them ideal for smaller cities that don't have the resources or infrastructure to implement more traditional systems. The only successful PRT system from this initiative was built on the Morgantown campus of West Virginia University and is still used daily to shuttle students to various destinations around the school [2]. Compared to a tracked system, the Morgantown PRT utilizes a relatively simple guideway for its small vehicles, consisting of just two concrete guides containing induction loops, providing power and system information. The main draw to this kind of system is to minimize cost and lower impact on city infrastructure. However, in the case of the the Morgantown PRT, the budget was greatly underestimated and costs such as track de-icing had to be factored in. For this reason, many other PRT systems were abandoned and replaced by more established technologies. PRT systems are the closest match to Robot Highway as they operate autonomously transporting passengers to destinations. This project is different in that the vehicles operate on one fixed path and traffic rules or collisions are not a issue. 

Figure 3: The Google Car

Another solution to the issue at hand is a project called the Google Car. The Google Car is a system in which a vehicle may operate autonomously using information from Google street view and artificial intelligence [3]. This project is currently under development. The Google Car has tested in many instances successfully including a test where the vehicle drove 1,000 miles without any human intervention. In addition, Google is lobbying for driver-less vehicle laws in the state of Nevada. Google Car, while addressing all the issues of Robot Highway, this project is beyond the scale of Robot Highway as it has no fixed highway.

Design goal

 The goal of the robot highway is to design a robot that will act as a taxi taking passengers around four stations on a highway network for robots in a 5x5 ft arena. The robot is designed to earn money as it delivers passengers and must obey traffic rules as it travels along the arena. 

Project Deliverables

The group delivered an autonomous robot taxi that could successfully transport passengers around 4 stations of a highway while simultaneously abiding the standard traffic rules. Along with this design the group delivered an instruction manual detailing the travel algorithms in which the robot operates, a thorough inventory of the individual components of the design, and a map of the highway.
 
The travel algorithm details how the robot taxi  has the ability to recognize and deliver passengers to the stations using  sensors that continually interpret the location of the robot in conjunction with the passengers' destinations. The inventory specifically outlines each piece used to create the design and the total cost of the robot as a result of the summation of components. This was then weighed against the projected total net cost made from bringing passengers to various stations. The manual also included a map of the highway in which the taxi operates shown in Figure 1 on this page. The group was able to provide an efficient and effective robot that could successfully operate around an automated highway.

  Project Schedule

Projected Budget 

The NXT Brick was the brain of robot, this part was integral to all other aspects of the design. This part had the highest cost, it ran all of the programming scripts written on Mindstorms and is connected to all sensors and motors that performed the various operations performed by the taxi. The motors have the same cost as the NXT Brick as they have the ability of controlling the motion of the robot. The sensors also had a high cost: color, light, and ultrasonic cost 20, 30, and 20 respectively. Without the sensors, the robot would be unable to detect its' location or interact with its surroundings. The sensors are fundamental in the robots navigation of the highway, passenger transport, detection of other robots, and all other types of interaction. The smallest cost was the various small and large parts that were used to build the structure of the robot. These parts constructed the frame of the robot, and connected all of the various motors and sensors to the brick of the robot. The net cost of all of these components in addition to the cost of the motors, sensors, and NXT brick brought the total cost to 234.