Click here to go
ABSTRACT
In the coming years of very sophisticated technology there is a need for implementing intelligence for a vehicle. That is it should go to the prescribed destination may be without the human interference .Our project is an attempt showing how this is possible.Here our project'IMPLEMENTING THE OBSATCLE AVIODANCE ALGORITHM' is implemented for a small test vehicle with some predefined assumptions. For a system to operate as an intelligent system it should have a very important component called the control system.So using the control system we can control the motion of the vehicle and direct it as the we wish. We have used a programmed microcontroller as the control system. The program loaded in it controls the actual motion of the vehicle.The basic principle here is that we sense if any obstacles present in the path of the vehicle or not and then proceed avoiding them by taking necessary turns and the compensating turns.We have used infrared sensors for this purpose.
Coming to the future scope of the project we can improve the present model to a very sophisticated one .We can directly feed the path and the speed for the vehicle i.e. in which it is expected to go.So that it can move from one place to the other by implementing for a huge vehicle.
1. INTRODUCTION
Embedded software solutions let us include obstacle avoidance ability various types of mobile systems, from toys to appliances to sophisticated autonomous robots.
The term obstacle avoidance describes a set of software techniques that allow mobile machines, such as housecleaning appliances and other types of robots, to adjust their trajectory according to their surroundings. Used in conjunction with distance measurement and motor control solutions, this can give autonomous products the reflexes to handle obstacles intelligently, even when those obstacles are in motion themselves.
Generally, an automated vehicle will require some means of detecting obstacles in its path, and a capability of maneuvering to avoid them. A variety of acoustic systems have been developed for this requirement over the past 20 years. These range from simple, single beam echo sounders which look ahead of the vehicle to more sophisticated multibeam sonars which can detect, track and classify an obstacle and then carry out one or more pre-programmed avoidance maneuvers.ISE has had experience with a number of these systems. In the ‘80s with Kongsberg Simrad Mesotech Ltd, they developed and tested an 8 beam system using Simrad Model 1071 echo sounders in our ARCS vehicle. This system was very successful and demonstrated the capability to detect and classify an obstacle and then perform a simple avoidance maneuver. This system was later upgraded to provide the ability to track objects including depth contours at a preset range. They also conducted experiments with fewer beams (transducers), and demonstrated that a fundamental capability could be met with much fewer beams. In the ‘90s, they integrated a Sonatech, Inc. 20 beam obstacle avoidance sonar which provided the vehicle with sufficient information to undertake sophisticated maneuvers .The level of sophistication required by an automated vehicle will be determined by the mission, and more specifically by both the frequency of encounter with obstacles and potential complexity of the situation.
Even though an automated vehicle is capable of fully autonomous operation, a well-designed system provides the operator with communications links to the vehicle whenever possible. At a minimum, a link is needed to start the mission after the vehicle has been launched. This contact confirms that all systems are ready and allows the operator to command the start with knowledge that the vehicle is fully functional . Communicating with the vehicle at the end of a mission is very important. It is important, however, to be able to assure that the vehicle is safely shut down and the link allows that to verify or to accomplish .
The obstacle avoidance algorithm is very complex in nature.So in order to simulate it we have mad certain assumptions.The main assumption here is that the vehicle to which the algorithm is to be implemented moves only in the forward direction and it avoids obstacles ahead of it.Even after avoiding the obstacles it has to move in its original or intended direction.That is it has to maintain its path irrespective of the obstacles in front of it .
Obstacle Avoidance and Motion Planning are both necessary components for any robotic task. Obstacle Avoidance refers to planning collision-free trajectories for robotic systems while Motion Planning refers to planning smooth motions for robotic systems.
Obstacle avoidance methods based on ultrasonic sensors must account for the sensors' shortcomings, such as inaccuracies, crosstalk, and spurious readings. So at a better level we have gone for using infrared sensors because of the fact that IR sensors are cheap and accurate .
2. ALGORITHM
The main intention of the algorithm is to get rid of the obstacles present in its intended path.The main point in the algorithm is that the vehicle turns to the other direction that is to the left or to the right to some angle so that it can find another route and moves in that direction till there is no obstacle and finally covers the angle it has made previously to avoid the obstacle. In this way the vehicle moves in its intended direction avoiding the obstacles on its way.
The algorithm simply takes the desired destination and finds the straight line path to that destination. Next it looks at where the vehicle will be by the next frame if it travels this path and it also looks at where the various obstacles will be at this future time frame. If there is a problem with the future position, it leads to getting tagged or is in a forbidden area, then a new destination point is found. This destination point is found by taking the original destination and slightly moving it. This is repeated until a destination point is found that is as close to the original as possible, yet results in the vehicle being in a safe location by the next frame.
3. WORKING
When the initialization that is when the power is switched on for the vehicle the microcontroller starts executing the program in its ROM as the signal at its EV pin is high . During the initialization appropriate values to the ports will be set up and the controller is made to wait for the start bit from the user .Until the user gives a start signal to the vehicle it sits in its position without any movement .As soon as the controller gets the wait signal from the user its starts its actual motion and implementing the algorithm .The flow chart for the algorithm is in the figure 2.1 .
The actual implementation of the algorithm is as follows .First of all it turns on the front sensor and then checks for any obstacles .If there is any obstacle in the front that sensed by the front sensor , it turns on the left sensor and the again checks for the obstacles .If there is an obstacle in the left it turns on the right sensor and then repeats the same checking . If still there is an obstacle in the right the vehicle comes to an under standing that there is an obstacle that is covering a total of one side and it is made to turn left and then implement the algorithm from the beginning .
If there is no obstacle in the right ,( that is there is an obstacle sensed for both front and the left ) it checks for the number of right turns it has taken ,if it not compensated with the direction it simply turns right moves forward and continues from the first point .
If there is no obstacle in the left ,( that is there is an obstacle sensed for the front ) it checks for the number of left turns it has taken , if its not compensated with the direction it simply takes a left turn ,moves forward and continues from the first point .
If there is no obstacle in the front it checks for the equality of the number of turns it has taken to the left and that to the right . If they are equal then the vehicle moves forward and continues from the first point . If the number of left turns is greater than the number for right turns it turns on the right sensor and checks for the obstacles if there are no obstacles it compensates the lefts it has taken . If the number of left turns is less than the number for right turns it turns on the left sensor and checks for the obstacles if there are no obstacles it compensates the rights it has taken .
In this way the vehicle moves forward avoiding the obstacles present in its path thus implementing the obstacle avoidance algorithm .
4. IMPLEMTING THE ALGORITHM FOR THE VEHICLE
Here the obstacle avoidance algorithm is applied to a small vehicle whose basic principle for running is DC motors .We have formulated the idea of the algorithm on the flowchart , but to get in to the real time trouble shootings we have to implement it to test vehicle . The main component of the test vehicle is the microcontroller . It is actually with this we implant the algorithm to the vehicle . When we implant is for vehicle we will be getting the exact picture how it can be applied to a huge vehicle and what are the difficulties with complex obstacles . Getting the real time picture for the vehicle we can easily make the modifications and make it lively
5. APPLICATIONS
Wany has developed very efficient navigation algorithms to avoid and bypass any obstacles, regardless of their color, material and of their shape and thinness (table, chair leg, …). Obstacle avoidance is one of the key navigation requirements for housecleaning autonomous mobile devices. Regarding thin obstacles, Wany uses triangulation and dead-reckoning-error algorithms to enhance the performance of the telemeters and obstacle avoidance software. This functionality is difficult to obtain because of the high price of precise telemeters (often ultrasonic in this case), Wany Robotics can use both market-available telemeters and our patented IR telemeters to perform the obstacle avoidance algorithms at a very low price.
EXAMPLES :
Obstacle avoidance is a key feature in any kind of robot , household appliance , PC peripheral , or toy that needs to move around autonomously , including:
• Vacuum cleaners , lawn mowers
• Toys cars , boats , tanks , animated animals
• Remote Web cameras, surveillance robots
• Personal robot assistants
• Vehicle security and advanced personal warning systems
KEY FEATURES :
•Obstacle avoidance technology gives mobile machines and robots life-like reflexes and allows them to navigate intelligently .
•Software is hardware independent (supports any environment sensor technology , including infrared , ultrasound , and video) .
•Solutions are adapted to requirements of the target system , and are completely scalable for use in any kind of mobile device .
•Supports systems with as little as 1 KB RAM and ROM .
•Scalable upwards from 8-bit, 4 MHz processors .
Click here to go