Robotics gets interesting… step by step

The Romford Robot plays snooker. For those unfamiliar with the game, snooker is essentially a two-dimensional activity that involves poking balls around a flat table-top surface with a stick (which is more commonly known as a ‘cue’) and manipulating the angles generated when they hit each other to steer them into holes or ‘pockets’ which are set at regular intervals around the table’s edges.

The Romford Robot has, over a period of twenty years, proven to be highly effective in calculating the speed, angle, and spin necessary to excel at the game consistently. All of this would be quite exciting were it not for the fact that the Romford Robot is actually the nickname of a human player.

cc  by  OliverC999

Steve Davis is a six-time world champion and was ranked as the world’s number one player for seven years in succession between 1983 and 1990. Still playing professionally today, Davis has nicknamed the Romford Robot to recognize his hometown – Romford – and the seemingly emotionless and error-free manner he played. For almost a decade, Davis singlehandedly made snooker betting a remarkably consistent investment.

Snooker is a game (some say sport) requiring a high degree of fine motor control and the practical mathematical ability to calculate the angles and speed, which snooker balls will realize once they are in a collision.

Enacted across the two dimensions of a flat tabletop, the challenge posed by this combination of calculation and precise kinetic execution has attracted the interest of robotics engineers. As implied by the former world champion’s nickname, the challenge was straightforward: is it possible to actually produce a snooker-playing robot capable of replicating Davis’s amazingly consistent level of performance?

In 2011 an engineering PhD student project involving Thomas Nierhoff, Omiros Kourankos and Sandra Hirche of the Technische Universität München (Munich University of Technology) developed a plausible candidate in the form of a two-armed ‘Snooker Bot’ which used its 7-DOF arms, pre-programmed pathfinder software and an overhead camera to replicate the actions of a human snooker player. Snooker Bot was capable of potting five balls in succession from the positions achieved after each successive shot to a threshold of 80% accuracy.

A five-ball sequence would not impress the Romford Robot, who repeatedly showed himself capable of reaching the highest possible score in snooker 147.  The snooker maximum requires the consecutive potting of 36 balls – all without causing any other balls to drop into any of the six pockets inadvertently.

The TUM researchers are not alone in having sought to replicate humans’ ability to master snooker’s intricacies. In the US, a Silicon Valley start-up going by Willow Garage’s name developed its own version of a pool-playing robot (pool is a snooker derivative and is directly comparable in terms of the demands it presents. More on pool vs snooker.

Using laser technology and high-resolution cameras, the PR2 (Personal Robot2), the team could customize FastFiz, an open-source pool physics program developed by Alon Altman, and combine it with a specially designed grip and ‘bridge.’ It is perhaps a reflection of the PR2’s ‘success’ that its developers quickly adapted it to fetch drinks from the fridge as an alternative to playing pool.

The development of sporting robotics is a perennial feature of trade shows – the Omron table tennis machine is perhaps the most famous – but the full complexity of the range of human perception, action, and reaction remains a fascinating challenge.

The possibilities explored in ‘Ex Machina’ may offer tantalizing glimpses of what might be possible, but in the real world, things are inevitably less glamourous. That is not to say that the challenge of bringing scientific fact up to science fiction’s speed does not remain compelling. The various robotics on show at CEATEC 2014 undoubtedly represent steps in that direction.

ccbysa   by  TAKA@P.P.R.S

In addition to the Romford Robot, one of Steve Davis’ other nicknames was ‘Interesting’ – as in ‘Steve Interesting Davis.’ It was meant semi-ironically, but for those keen to find a way to mechanize the seemingly automatic accuracy that Davis brought to the snooker table, things are getting more interesting all the time.  Much like Davis building a break by methodically calculating one shot at a time, robotics’ development continues to build as a series of seemingly isolated and unspectacular (if not uninteresting) steps.

Leave a Reply