What counts as a service robot
The word “service robot” is used loosely. International standards bodies have described it as a robot that performs useful tasks for people or equipment, excluding industrial automation applications. In practice it covers a wide range of machines: robotic vacuums, hospital delivery carts, warehouse shelf-movers, lawn mowers, window cleaners, security patrols, shop assistants and social or companion robots. What they share is that they work in places designed for people, not in cages on a factory floor. This article follows how that second wave of robotics developed after the factory arms of the 1960s and 1970s.
Early ideas and obstacles
By the 1980s the commercial success of factory robots encouraged speculation that the next market would be homes and offices. The idea was attractive, and the difficulty was underestimated. A factory is structured and predictable, while a living room or a hospital corridor is cluttered and full of people. The sensors, computers and software of the time could not reliably cope with it. Many robots that were shown at fairs in the 1980s, including entertaining home butlers, remained demonstrations.
Hospitals and delivery
One of the first real service markets was hospital logistics. Joseph Engelberger, who had helped found the factory robot business, started a company in the 1980s to build a hospital delivery robot, which was sold as HelpMate. It carried meals, medications, laboratory samples and supplies along hospital corridors, using sensors and stored maps to navigate and to take elevators. Later autonomous carts for hospitals and hotels built on the same idea.
Hospitals suited the technology because the tasks are repetitive, the routes are fixed and there is a staffing reason to automate. Even so, adoption was slow and gradual, and reliability around crowds was a recurring problem.
Cleaning at home
The household robot that finally reached mass markets was a vacuum cleaner. In the early 2000s, several companies introduced small, low-cost machines that wandered across floors, steered around obstacles and returned to a charging base. Their success came from setting a modest goal: one job, a flat surface and tolerance for imperfect coverage. Later models added better mapping with cameras or lasers, and similar designs followed for mopping, lawn mowing and pool cleaning. These robots, and the broader family of vacuum cleaners they belong to, showed that consumers would accept robots when the task was clear and the price was reasonable.
Humanoids, pets and companions
A separate line of work produced robots designed to look or behave like animals and people. Sony’s dog-shaped entertainment robot was sold from 1999, and Honda’s humanoid ASIMO, shown in 2000 after years of research on walking, became a well-known symbol of the field. These machines drew large public attention but were not widely useful for chores. In the 2010s, several companies introduced social robots designed for greeting people in shops or banks, for companionship or for education. Studies and reports on their long-term usefulness are mixed, and some products were discontinued after limited sales.
Companion robots are also being studied for older people and for children with learning needs. Research results so far are varied, and there are ethical questions about attachment and privacy, which developers and researchers continue to discuss.
Warehouses, shops and public spaces
One of the most commercially significant service uses has been in logistics. Warehouse systems in which mobile robots bring shelves or bins to human pickers were developed in the 2000s, and Amazon’s purchase of the company behind one such system in 2012 helped bring them to wide attention. Delivery robots on sidewalks, floor-cleaning machines in stores, and inventory scanners have followed. Most of these are used in managed environments with maps, rules and staff who can help when something goes wrong.
What made service robots possible
Several technical improvements underpinned the growth: cheaper sensors, particularly laser range finders and cameras; mapping and navigation software that lets a robot build a map and locate itself within it; smaller electric motors; and lithium-ion batteries that allow hours of use. More recently, advances in machine learning have improved how robots recognize objects and speech, although they have not removed the need for human supervision in many settings.
Lessons from the history
The service robot story differs from the factory story. Factory robots succeeded by changing the environment to fit the robot. Service robots must fit into spaces made for people, so the successful ones have usually been narrow in purpose. They also depend on customers accepting robots near them. The general-purpose household helper that appeared in early forecasts has not yet arrived, and projections of when it might arrive are uncertain.