AI Problem:: Robot Navigation System

You might also like

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 5

AI Problem :

Robot Navigation
System.

Shubham goyal
2015ucp1558
Abstraction:
1. Goal of the robot navigation system is to provide better logistics
mechanism in factories.
2. Find best solution to route the robot through factory mazes and
supply the inventories to various junctions and return to initial position.
3. Algorithm can also be used as intelligent trolley to load items in
shopping malls.

Problem Description:
1. Maze is given as polygon obstacles . Maze can be a rectangular grid
that can be generated by pseudo random number generation
algorithm.
2. Various junctions are situated in the maze where the robot has to go
and return to the initial state.
3. Problem can be divided as :
a. Route finding
b. Obstacle avoidance
c. Travelling salesman problem

1
Problem Formulation:
1. Input Module will process the environment through cameras, is there
any obstacle in front of it or not.
2. It will keep a database of each junction visited and will keep track of
the path.
3. Speed of the robot is controlled by range of obstacles from itself.
4. To process the route it is using backtracking Depth first search
algorithm that will generate a path from initial node to next junction.

● Initial State:
○ { Visited[NULL] , Unvisited[N] , Pathlist[NULL] , PATHCOST, 2-
DImage[]}

Here, Pathlist -> stack of visited nodes .

Visited[] -> array to store visited junctions.

Unvisited[] -> array to store unvisited junctions.

PATHCOST -> variable to store cost associated with path.

● States :

2
○ { Visited[ junctions] , Unvisited[junctions], Pathlist[a,b,c,...], 2-
Dimage[] }

● Action :
○ moveleft(), moveright(), movefront() .
○ if( junction) : stop() && wait(20 sec) && keepsense()
○ Visited[]++ , Unvisited[]--
○ if( !workload ) : keep visit to store information about paths to
calculate heuristics().
○ PATHCOST + = distance
● Goal Test:
○ State==Initial state && workload==NULL

AGENT MODELING:

● Performance:
○ Pathcost of the travel.
○ Energy consumption during travel (Travelling salesman problem
implementation ).
○ Obstacle collision frequency.
○ Junction order missing or placed database.

● Environment:
○ Obstacle graph generation(automated grid generation).

● Actuators:

3
○ Image Processing.
○ Speed controller.

● Sensors:
○ Front camera to make a picture.

Agent Architecture:
It is a goal based agent which is focused to maximize the performance
of routing.

Front
Image processing cam E
n
v
I
Obstacle graph for dfs. r
o
n
m
e
Travelling salesman algo n
t

Action to actuators. action


and
speed

You might also like