‘The laws of robotics will always remain relevant’
Published on : Monday 30-11--0001
Are we in for a Robotic ‘takeover’ of manufacturing?
The need for a complete takeover of manufacturing by robots has been there since a long time. However majority of the robotics industries have confined to certain regulatory standards and hence disruptive innovations in industrial robots have not really happened. Conventional robots have not been safe for co-working with humans. Specific discrete manufacturing processes such as assembly of parts have been fully automated using special purpose machines. But operations such as material transfers have been still aided by humans. One of the prime reasons why automated material transfers over long distances in a factory setup is not seen everywhere is the high capital cost. Quality control has been highly human resource intensive in most of the industries.
But certainly there is a silver lining. Certain disruptive robotics technologies developed by companies like Rethink Robotics which have exploited complaint mechanisms with force control, stereo-vision, enhanced learning and smart collaborative capabilities would be the first step towards faster adoption of robotics in manufacturing. Smart Robots such as Baxter and Sawyer and few others can be trained without programming, they can work in human environments, and their stereo vision can make them perform any kind of operations due high dexterity. AGVs and mobile robots with smart manipulators would be able to perform unstructured material transfers. In my opinion, a collaboration of special purpose machines, dedicated industrial robots, smart robots and mobile manipulators is very important for complete automation of a manufacturing setup. When these robots are connected and synchronised over real time industrial networks and supervised by central computers running intelligent algorithms for collaborative processes, statistical process control, process enhancements, fault predictions, etc., going beyond the Industry 4.0, can turn it into an intelligent robotic manufacturing setup. This day is not far and economics of adoption would certainly look attractive.
Will the new generation of robots make human presence redundant on the shop floor and restrict it to a supervisory role?
That is the ultimate goal of intelligent robotics. The humans can play a better role in designing process enhancements and also teaching the smart robots what to do, and maintenance of the machines. The humans will have more white collared jobs such as data collection, smart management of the machines, analysis of the trends and predictive maintenance of the machines so as to reduce downtimes. Making smarter process measurements and enhancing the processes to improve productivity, energy efficiency, and material efficiency while striving to achieve zero manufacturing defects targets will be the prime responsibility of the humans.
On the other hand, do we need more and more human beings for R&D on robots and controls?
While intelligent robotics aims at reducing the human presence in manufacturing setups, the need for qualified manpower to work on development, programming/teaching, commissioning and maintenance of intelligent robotics systems is going to increase drastically. The human intellect is being continuously challenged to develop more intelligent robots. Research and development in advanced electromechanical actuation systems, series-parallel kinematic linkages, stereo-vision and other sensing technologies, motion control and task planning, intelligent co-ordination between multiple robots and humans, speech understanding, high speed real time networking and advanced visualisation is key to the development of intelligent robotics.
Are Isaac Asimov’s laws of robotics relevant anymore?
The laws of robotics will always remain relevant as the ultimate goal is not to harm a human being, except for defense applications. The extent of human orders to the robots will change but the essence of the second law will still remain relevant. The third law brings in an interesting scenario wherein the robot is expected to perform operations without hurting itself, which means it has to maintain its integrity and health. So in case of malfunction of a certain module, the robot will be expected to replace the module by itself and gain self recovery mechanisms, thus taking the autonomy to the next level.
Can robots replicate all the intelligence of a human being?
It has taken millions of years for the super intelligence of a human to evolve, so on a lighter note, aren’t we expecting a bit too much, for the robots to catch up with human intelligence? The research and development efforts in robotics have been primarily focused towards improving the dexterity of the robots, vision and environment understanding capabilities, locomotion, miniaturisation, multi-robot co-ordination and human-robot interaction. These areas are directly influencing the productivity, usability and relevance of the robots in industrial, and service sectors while minimising the costs. Thanks to the semiconductor and microprocessors industry, the computing capabilities of the robots have gone up exponentially and the networking and connectivity to internet has become a de facto. And thanks to the IoT and the big data revolution, data mining, machine learning and artificial intelligence have become household terminologies. The day is not far when advanced intelligence and access to the world information would get integrated with the advanced robotics systems; and therefore making the robots close to a human or in some cases better than humans.
Captions
Photo1: Rethink Robotic’s collaborative robot Sawyer at a work station. Photo Credit: Rethink Rebotics
Photo2: Robots are capable of locating workpieces in a crate. Photo credit: Fraunhofer Institute.
Synedyne Systems
Bangalore