Connecting people to machines anywhere in the world
M. Rokonuzzaman | Published:
December 06, 2015 22:46:49
October 23, 2017 03:00:00
Internet of Things (IoT) is now a big thing - like the way we heard about .com in the late 90s. The IoT refers to the networking of physical objects or machines through the use of embedded sensors, actuators, and other devices that can collect or transmit information about the objects. Although examples in the consumer market include smart watches, fitness bands, and home-security systems, but the B2B (business to business) market include sensor-embedded semi-autonomous production equipment. By many estimates, by 2020, there will be between 40 to 50 billion connected devices - from homes, cars, clothing and factories - that will transform everyday objects into powerful data-acquiring machines, with the promise that they will positively change the way business is conducted. The question is: whether data accumulation capability alone will enable IoT make our world a safer, more efficient, and more prosperous place. Until and unless consumers see money in it, will interface between Internet and bunch of built-in data gathering sensors produce a fraction of trillion-dollar economic benefit, which is touted for? Will data-centric approach likely end up in repeating the .com disaster of the late 90s? To meet the ultimate goal of IoT to increase operational efficiency, power new business models, and improve quality of life-IoT will likely be more than empowering machines to collect massive amount of data.
All the machines produced by the Industrial Age need people in the loop, whether it is turning on/off light bulbs when needed or driving a car. Without the integration of sensing, perception, and basic decision making capability, to be easily done by even a high school dropout from developing countries, most advanced driving machine of BMW or Ferrai is a useless piece of metal object. Over the last 50 years, there have been significant efforts to develop fully autonomous machines--commonly known as robots. These research efforts, including Honda's ASIMO project, have taught us a very important lesson: it's far more complex than initially thought to be to add basic perception and decision making capabilities to these machines to make them do meaningful job in real world, without requiring human operators in the loop. Although, initially it was perceived to be sensors, but tremendous growth of capability of cameras has proved that complexity is somewhere else. Experimentation of computational equivalent neural networks or fuzzy logic has taught us a lesson that basically, we have no clue to computationally perceive enormous amount of data, produced by cheap sensors, to understand simple variations of real-life situations. Therefore, it may not be unfair to comment that enormous amount of data generation capability of these machines, coupled with Internet interface, will have little value to add -- to encourage consumers to pay a fraction of trillion dollar for that.
In advanced economies, many of the jobs most resistant to automation are those with the least economic value. Just consider the diversity of tasks, unpredictable terrains, and specialised tools that a landscaper confronts in a single day. No robot is intelligent enough to perform this $8-an-hour work. This opens the opportunity of robots remotely controlled by a low-wage foreign worker. Speed of Internet connections, cost, and the latency involved in long-distance communication have so far impeded progress toward the "avatarisation" of the economy. But emergence of low-cost, high bandwidth Internet connection and standards of IoT are opening the opportunity of connecting millions of workers from Bangladesh and other developing countries to robotic avatars in Japan, USA and many other advanced economies to carry out non-routine work. It's likely that IoT will evolve to encompass next generation methods and procedures such as "teleoperation" (operation of a machine at a distance), tele-robotics, and other areas that rely upon interface and control of real objects or machines by remote operators.
By taking the advantage of IoT, the teleoperation technology developed for space operations, and also for other special terrestrial applications, may be very useful add-on to these sensory-rich industrial products, connected to Internet. In addition to the capability of producing data to make the work environment visible from distance over the Internet, these machines will have semi-autonomous capability to create economic viability to benefit from IoT. Semi-autonomous capability of a lawn mower may be like: 1. Move left, 2. Move right, 3. Keep mowing along the straight-line until you detect the boarder, 4. Go to the corner of the lawn, or 5. Get ready for mowing, etc. Once these semi-autonomous capabilities are built in these machines connected with a rich set of sensors and Internet - also there should be a set of cameras at different locations to give seamless view of the work space to a distant viewer over the Internet - someone from Bangladesh or other developing countries will be remotely supervising these machines to do the lawn mowing. Payment of less than USD3/hour will be certainly huge saving for the lawn owners of North America, Europe or Japan, but that could be good pay for remote operators. Due to semi-autonomous nature, multiple of these machines could be even deployed under the supervision of a single operator, resulting in further lowering the cost of operation. It appears that such semi-autonomous capability to allow teleoperation of these devices from developing countries are going to generate economic value to both remote service providers and consumers of advanced economies. It seems that companies which develop consumer product would now need to adapt Space tele-robotics for mass-scale commercial operation to start a new wave of wealth creation through IoT - of course, with the integration of perception and decision making capabilities of millions of youth of developing countries.
To take advantage of this emerging opportunity, connecting people to machines anywhere in the world, we need to have super high-speed broadband connectivity - at gigabit per second (Gbps) speed - to ensure virtual presence of workers in the remote work sites. Due to high volume data-centric nature, cost of data connecting millions of Bangladeshi youths to globally distributed IoTs should come down to a small fraction of current price. The price of data should not be more than 5.0 per cent of income level, as outlined by the UN, of these future workers to enable them to remain virtually present in remote work space in operating future sensor-rich machines. To make progress along this line, policy makers should create possibilities of profitable competition for connecting millions of Bangladeshi workers, primarily from rural areas, to remote machines through super high-speed broadband connection at a very cheap rate. The capitalisation of this opportunity may end up in much higher bandwidth demand in rural Bangladesh than cities -- quite contrary to our belief though. The evolution of IoT technology and economic viability are at a tipping point to resonate to create large-scale business opportunities to transform the way the world works with machines of old Industrial Age. In this new Industrial Age, these dumb machines will be augmented with remote intelligence to amplify the benefit of industrial economy by many folds as well as to reduce downsides. To avail this emerging opportunity, we need to connect our villages, towns and cities with super high speed, Gbps, broadband connections - creating a new invisible way, as a replacement of highway or airway, to connect our future millions of job-seekers to billions of semi-autonomous smart machines spread across the world.
M. Rokonuzzaman, Ph.D. is Professor, Department of Electrical and Computer Engineering at North South University.