Optic Flow in Flying Robots

There has been much research in the field of optic flow, in particular the navigation in insects and robots. The vast majority of work in optic flow based vision, emphasizes on the different algorithm used to enhance the image processing. Here, I discuss the advantages and disadvantages of applying optic flow navigation, in particular flying robots rather general, and note that we need to consider different parameters when the robot is flying. Consequently, I will discuss some of the practical issues in this field.

Optic flow can be defined as the information of the environment or situation which is flown into animals or machines controlling system with the help of vision is called as Optic flow. According to Qiu (2000) “Optical flow is usually descriptively defined as the apparent motion of the brightness pattern in the image plane or an observer's retina resulting from the relative motion of the observer to the scene or the moving objects”.

Optic flow Mechanism in Insects/bees

Insects use optic flow mechanism extensively in all their activities and shows behaviour accordingly. Eyes plays very important role in insects and hence optic flow. Insect’s eyes are composed of compound eyes and these eyes have small eyes called ommatidium. They have large number of photoreceptors. Few of them, around 6 to 7 are connected to each ommatidium. Ommatidium are placed in such a way that, insect can have panoramic environmental view (Srinivasan and Zhang 2004).

Insects cannot find the distance of an object just by looking at it and their eyes posses very less range when compared to human eyes. This made insects to adapt or evolve different visual techniques (Srinivasan and Zhang 2004). We can re-create some of these things of bees and put them into novel kind of things. These alternate visual techniques of insects had made researchers to find different ways of solving problem in different fields. One of such largely used field is Robotics and its sub fields, autonomously navigating robot, study of insects behaviour, aeronautical, military applications and many more.

Advantages:

• In Narrow Gaps - Insect finds how rapidly the image of the world moves on their eyes and find how far they are located and navigate in it. Insects actually translate speed of image motion into range of object. The above optical flow technique can be used in flying robots which helps to fly in the narrow gap. The advantage is that they need not find the distance between the walls or on either side of the wall.

• Controlling Flight Speed - Bees control their flight speed by observing how rapidly environmental picture moving pass them and keeping that image velocity in eye constant.The centering mechanism and flight speed controller is done by the same system. That is, the system that controls the flight speed also does the centering mechanism, hence you do not require two different systems to carry out two different techniques.

• Collision Avoidance - We all know that whenever an object approaches the eye, the image of the object in the eyes increases. The same mechanism is used by insects to avoid collision. This technique can be implemented in flying robots.

• Smooth Landing

• Goal Directed Navigation

• In General

1. The robots do not need to know how fast it is travelling.

2. The robots do not need to know how fast it is approaching the ground.

3. The robots do not need to know how high it is above the ground.

4. The robots do not need to have different systems to do the above task, only one system can do all the task.

Disadvantages:

• Since it is the one system that takes care of all the properties like flight speed, centering and others as explained in Section 4.1, if this single system fails in the flying robot, then entire robot fails to perform a major function called ‘Flying’.

• Since the real world is 3D, it is hard to analyse the image and image variation to world motion. The spinning sphere image does not change regardless the sphere’s motion, similarly if it is a uniform plane as wall, then the camera motion (vision device) may not bring on any change in the image.

• It has been observed that applying vision in real task has two major limits, they are, robustness in real world, uncontrolled environment and the computing power required for real-time performance. They are very sore to noise and inaccuracy.

• Another unidentified situation is transparent medium like “Glass”, where it is difficult for robots to identify the medium using optic flow, there is a risk of collision. Since normal cameras has no capability to distinguishing between transparent medium and atmosphere.

• Since we use optic flow mechanism in our flying robots, therefore ‘vision device’ used in such case plays a very important role. Hence precautions has to be taken so that the lens of the vision device is not blurred due to weather condition in nature or natural calamities. The flying robot has to work in any weather condition.

Insects are the best field to explore the optic flow, in particular for flying robots.All the above advantages and disadvantages shows that how complicated the visual system in bees is. According to Webb (2006) many consequences remain unresolved like, the insect flying straight in front and sensors pointing down, then the ventral optic flow will be deformed if the insects rolls and pitches, control mechanism many are undiscovered.