State Switching Kalman Filter (SSKF)

Visual tracking has always been an interesting field. There are many filtering algorithms depending on the the process model and the measurement model of the object being tracked. Kalman filter is the most popular filter used to track objects which have a linear motion. Extended Kalman filter(EKF) and Particle filter are used to track objects that have a non-linear trajectory. Two years back at the Mechatronics lab in Clemson University we faced a unique problem. Dr Hoover had designed a workcell with a puma robot and visual sensing for the robot was provided using a 6 camera network. The idea behind the workcell was to develop a robotic arm which would adapt to changes in its environment.

Industrial manipulator nowadays are very efficient and precise but they are blind. When a robotic arm inserts a chip into a board, it assumes that the board would be at the specified position at the specified time. Any change in the environment would be disastrous. The workcell was designed so as to come up with a system that would react to changes like humans do. The first task was to track and grasp a foam ball moving in a random manner over an air conveyor. The air conveyor surrounded the staubli robot and 4 fans at its corners caused the random motion. The average speed of the ball was close to 100cm/sec. An occupancy map of the worspace was developed by fusing the 6 images from the cameras and by background subtraction raw measurements of the ball's position were fed to the kalman filter which filtered the noisy measurements and gave an estimate to the robot which extrapolated and picked up the ball. Everything worked well except the Kalman filter being a linear filter would take a long time to latch back on after the ball has bounced off a wall. Here a bounce event is a temporary non linear event and using an EKF dosent help matters much. Particle filters are computationally expensive so we decided to design a filter that would be robust and would be based on the kalman filter.

We decided to use the correlation coefficient of the measurements to detect a bounce event and used least squares approach using the most recent measurements to latch the filter back to the true state a lot faster than a normal kalman filter(KF). We experimented on this filter using different bounce models, measurement noise and different values of forgetting factor (forgetting factor in the kalman filter is the amount of belief in the measurements - more noisy measurement means KF will have lesser belief and vice versa). We found out that the SSKF performs better when the measurements are noisy and forgetting factor tuning is not necessary unlike the KF.

This algorithm is simple and at the same time very effective. Hawkeye Technologies which tracks the cricket ball to give the television viewers the projected trajectory of the ball for analyzing LBW decisions would find this useful.

Disruptive technologies

"Disruptive technologies" has become one of the buzz-words of late. Its a technology thats disruptive in nature i.e. causes a tubulence in the market for the existing technology. At times this can drive up the market to a new high of expectation resulting in a raised standards. But at its worse it can spell doom for the existing market players.

I recently read an interesting article about Disruptive technologies/ innovations (http://www4.gartner.com/research/fellows/asset_93329_1176.jsp). Its a discussion between Howard Dresner, an eminent technology writer, and Dr. Clayton Christensen, a Harvard Business School professor and the preeminent expert in the area of disruptive innovation and business strategy.

Clayton has a different take on what qualifies to be called disruptive. As per him, "a disruptive innovation brings to market a product not as good as the products in the current market, and so it cannot be sold to the mainstream customers. But it is simple and it is more affordable. It takes root in an undemanding portion of the market, then improves from that simple beginning to intercept with the needs of customers in the mainstream later."

I beg to disagree.

I dont quite think that the new technology has to be more affordable to be disruptive. It can charge a premium for the value it adds and still cause turbulence in the market.

Lets take the example of the Audio CD. When introduced it was quite expensive than the prevalent audio cassette. (And still continues to be, even though the advancements and experience have resulted in considerable price reduction) It charged a premium for the increased customer experience. But it was disruptive enough even though less affordable and eventually dislodged the audio cassette from its prime position of being the preferred audio playback format.

Another such disruptive innovation is Open Source Technologies...and the topic of my next blog on technology. Watch this space! :)

The Promise of Swarms

Have you ever observed ants? They are so simple creatures but show a very complex collective social behavior. They are able to find food hundreds of meters away, build mounds thousands of times their own height, form complex hunting strategies etc. Each ant follows just a few simple rules that are based on their ability to track scent but the resultant task of these simple ants is mind boggling.
This collective behavior, often known as Swarm Intelligence, is found among different species in nature and has led researchers to study the scope, limitations and potential applications in fields like robotics, search, navigation and collaborative task handling.
Out of the many applications of collaborative swarm algorithms, Swarm Robotics seems to be an interesting one. Similar to ant behavior, imagine a number of very simple machines following simple rules but are able to do a very complicated robust search operation in hostile conditions. DARPA has been actively investing in Swarm Robotics for a similar reason. People in universities like MIT and CMU are studying the theoretical and practical capabilities and limitations of existing swarm-based approaches for robotics research applications.
Annual robotics competitions like Robosoccer are a few small pointers to the growing interest in this field.
These swarm based approaches have significant advantages over centralized approaches. They are evolutionary, they degrade gracefully and most importantly they are robust! No doubt swarm-based approaches are going to be building blocks of next generation technology.

This reminds me of a saying by a famous economist Kevin Kelly

"The surest way to smartness is through massive dumbness!".

...And inspired by this I write a poem

"Look at the ants
and look at the bees.
Each one is dumb
but everyone has the keys.
The keys to behave
however simple it might be.
But together they are
invincible as you can see."