'Rules are there to be followed' and here, survival demands strict adherence to this basic rule. But for many, mere survival isn't enough. Chances need to be taken and if these pay off, it can take you to greater heights. Thanks to the laws of nature, a great fall is inevitable if u lose control at these heights. A less dangerous option is to explore the domain of the unknown and emerge unscathed. Ultimately, the golden rule that can take you to the pinnacle is, "Opportunities are there for the taking - when you spot one, grab it".
Build an automated robot that can traverse a complex arena consisting of bi-lane roads with obstacles (static vehicles). Computer control or on-board processing may be used to aid the robot. The robot should be equipped with on-board camera(s) (wired or wireless) for imaging, and should steer itself through the arena, avoiding the obstacles.
You can also include optional features like tracking a dynamic car, enabling the vehicle to follow directions based on road signs, crossing bridges with bends, navigating through tunnels and moving in untrained environment. (Successful implementation of these features will fetch more points)
The robot must be capable of basic navigation (explained below) compulsorily. The participants may also implement one or more of the optional features (explained below) as they wish, which will fetch them more points upon successful implementation.
- The robot has to navigate based on image processing methods only. Usage of other sensors or any other type of navigation is not allowed.
- The arena will consist of roads which are separated into two lanes by a solid line. The lanes are referred below as left and right lanes from robot’s point of view.
- Obstacles (static vehicles) on each lane are placed randomly and may be changed for each participant.
- The robot will be initially placed anywhere in the arena in any lane.
- The robot should always travel in its lane only (suppose if the robot is currently in the left lane, it should stay in the same lane, even after turning) until it encounters an obstacle and should shift its lane in that case without colliding with the obstacle.
- The robot is allowed to change its lane at most 15cm from the obstacle only.
- The roads may be straight, curved (smooth or steep curves), right angled and may also have junctions and dead ends.
- In case of dead ends (the roads may end abruptly without indication) the robot is expected to reverse and take the other lane.
- If the participant chooses to implement only basic navigation mode, in case of road junctions, the robot has to turn left if it is in the left lane and right if it is in the right lane (so that it remains in same lane even after turning).
ROAD-SIGNS DETECTION (optional)
- As the robot is travelling in basic navigation mode, it may encounter more than one road (junctions). The road signs are to be used to decide the path to be taken.
- The road signs are marked on the face of cuboids and placed in front of the lane in which the robot is expected to be while encountering junctions (there will be no road sign in front of the other lane).
- The road sign is placed at a distance of 5cm outside the path.
- Road signs like left arrow, right arrow, square or triangle are used.
- The robot has to indicate the detection of each road sign before turning, using any method (which must be intimated to the event coordinator before the run).
- The directions represented by the road sign are as specified below
- Left arrow or triangle -- left turn
- Right arrow or square -- right turn
- If the robot encounters a road junction it has to check for the road sign in front of its lane.
- Road sign detected – it has to turn left or right accordingly and continue in the same lane in which the robot was, before entering the junction.
- Road sign not present – robot has to continue in its own lane as mentioned in the basic navigation mode (rule no 9).
- The starting and the ending of bridge will be an extension of any one of the lanes only.
- The lane in the bridges alone will have a solid line along the centre to prevent the robot from falling off the bridge. (Notice that this line is not a continuation of lane separation line).
- The bridges will have inclined surface initially till it rises to a maximum height of 10cm followed by a plane surface sufficient to accommodate the robot and then declines to ground level (same slope as inclined surface) at 90 o either to the left or right only.
- The main aim of this feature is to make sure the processing delay of the robot is minimized to avoid its freefall along the slope.
- Each lane will contain tunnels (side by side), only in straight roads.
- The 2 tunnels will be of different heights.
- The robot should find out the tunnel of greater height and pass through it.
- The robot is not tested for its capability to navigate inside the tunnel (dark space) using night vision processing.
- Irrespective of height of the robot the robot must enter the tunnel of greater height only.
HOT PURSUIT (optional)
- The car to be chased will be parked somewhere in the arena initially before the run.
- This car can be differentiated from the obstacles present on the road.
- As soon as the robot gets behind the parked car, it should stop. Only after the car starts moving should the robot chase it. As soon as it starts chasing it need not follow any lane rules. It can move randomly (as if there were no roads present) avoiding static obstacles and road sign boxes alone.
- If the car is made static, the robot is expected to halt within a circle of radius 75cm with the car as centre. In the halt position the robot is expected to be pointing towards the car (at least a small part of the car should be visible when seen from on board camera).
- The robot is tested until the event coordinator gets convinced that the robot is efficiently tracking the car (the car is moved and stopped at random intervals).
FACE THE WORLD (optional)
- Participants may come up with a solution for real-time navigation in untrained environment (plane surfaces only), avoiding obstacles of unknown nature (any shape, size, and colour).
- This is an optional feature for which the robot will not be tested in the specified arena. The robot is placed in the centre of a virtual circle and is expected to come out of it avoiding obstacles.
- Implementation of this feature without basic navigation capability is not entertained.
- Basic navigation - should be implemented compulsorily.
- Road sign detection, Bridges, Tunnels, Hot pursuit, Face the world - optional features.
- One or more of the optional features that the participant wants to be tested should be intimated before the run. Those features will be tested on a single run at different places of arena.
- Separate testing of each optional feature is allowed (only if it fails in the single run) with some negative marks.
- The arena will be a square of size 6.0 m x 6.0 m with a black coloured border of thickness 2.5cm at a distance of 20cm from the outer edge of the arena.
- The entire arena will have a white background.
- Roads contain 2 lanes each of width 30cm coloured black. They also contain a solid line of width 5cm coloured white along centre of the road.
- Stop zone is marked by a rectangle of size 30cm x 60cm coloured blue on the road surface (start zone has no indication and will be chosen by the event coordinator anywhere in the arena).
- The static obstacles are cuboids of dimension 30cm x 15cm x 30cm coloured red completely.
- The road signs are coloured green on white background and are marked on the face of cuboids of dimension 30cm x 15cm x 30cm.
- Road signs include squares of dimension 20cm x 20cm, triangles of base 20cm and height 20cm, arrows consisting of a triangle of base 10cm and height 5cm attached to a smaller edge of rectangle of size 15cm x 5cm.
- The car used for hot pursuit can be of any size from 10cm x 10cm x10cm to 30cm x 30cm x 30cm and coloured red in upper half and green in bottom half.
- The sample arena and dimensions of bridge and tunnel are explained below diagrammatically. Unlike plane roads the bridges alone will have a solid guiding line of 2cm thickness coloured white along the centre to prevent the robot falling off the bridge.
|Bridge||Hot pursuit Car Model|
- All dimensions are in cm (not to scale) and accurate to ± 5%.
- The sample arena shown here is without any obstacles and optional features.
- Segment of the road coloured blue denotes the end zone.
- The robot is allowed to have maximum dimensions of 20cm x 20cm x 30cm (L x B x H).
- The robot can be controlled through a computer (wired connection or wireless transfer) or by onboard processing.
- In case of using wired transfer the participant must ensure that no damage is caused to the arena due to the wires (such as disturbing the obstacle or car).
- If wired transfer is used, the wire, and only the wire, can project beyond the robot sizing specifications.
- Participants are not allowed to use more than 2 cameras on board (2 preferable).
- There will be a preliminary round for all participants.
- Any part of the arena (laid for finals) will be used for prelims.
- Only those participants who clear the prelims will be able to compete in the finals.
- A team may consist of a maximum of 5 members. Students can be from different institutions and departments.
- Participants have to bring their ID cards (issued by their college), they will be disqualified otherwise.
- All participants who clear the preliminary round will be awarded certificates of participation.
- Any doubts or ambiguities in the problem statement should be clarified as early as possible directly by contacting us, no cribs during the event will be entertained.
Further details regarding computer and camera specifications, scoring details and change in any of the rules will be updated in Event Updates tab. Keep monitoring it.
EMAIL : firstname.lastname@example.org
MOBILE : +91 94860 82575
EVENT DATE : 22 nd January,2010
ARENA WILL BE DISPLAYED FROM : 21 st January,2010
To encourage use of image processing techniques for real-ti me problems among software geeks, knight rider is divided as 2 sections and knight rider- software is introduced.
KNIGHT RIDER - HARDWARE (accomplish chosen task with automated robot and camera assistance)
KNIGHT RIDER - SOFTWARE (display the results of processed image alone - robot not required)
RULES RELAXED - KNIGHT RIDER HARDWARE
- If preferred by participants any method other than image processing techniques can be used as a part or the whole to accomplish the task but image processing techniques are given higher weightage for scoring.
- It was specified basic navigation is compulsory. But now participants are allowed to implement any of the optional features even without basic navigation.
- There will be no dead ends in the path anywhere in the arena.
RULES - KNIGHT RIDER SOFTWARE
- Pick up your topic of interest in the problem statement (click here) and solve it using image processing. Same rules as Knight rider - hardware, except there is no need of a bot in software section.
- For software section, the camera will be manually placed in required parts of the arena. The participant's code has to capture image, process and display results accordingly.
- Apart from this if you think you are a pro in image processing, pick up your own real-time problem and solve using image processing techniques. You are allowed to compete in knight rider- software section only.
- The software and hardware are 2 seperate sections and are judged seperately.
- Even a single participant can participate in both sections.
- Separate prizes will be given for both sections.
- For any clarifications feel free to contact us.
Abstract Details DOC (25k)
Problem Statement PDF (162k)
Sample Arena Pictures ZIP (95k)
Simulation Videos ZIP (1.6M)
Main inspiration for the " KNIGHT RIDER " - DARPA Grand Challenge ( US Department of Defense initiative ) MP4 (22M)
- Onsite Programming Contest
- Network Warriors
- Online Programming Contest
- K MAD!!!
- Game AI Contest