|The robot revolution is gaining pace, changing every industry it enters. Robots can automate tasks, lift heavy objects more quickly and safely, repeatedly tighten bolts to the optimal torque, and work with no breaks / Photo by: besjunior via 123RF|
The robot revolution is gaining pace, changing every industry it enters. Robots can automate tasks, lift heavy objects more quickly and safely, repeatedly tighten bolts to the optimal torque, and work with no breaks. All of these, they can simultaneously do while filling a shrinking workforce in many industries. But these robots would not be as sophisticated as they are now without the help of artificial intelligence. AI gives robots the computer vision to navigate, sense, and calculate their reactions accordingly.
AI has the power to give life to robots and empower them to make their decisions on their own. Many industries are increasingly investing in AI-powered robots because there’s so much potential in them that needs to be released. The 2018 World Robotics Report released by the International Federation of Robotics (IFR) showed that the annual global sales during that year reached $16.5 billion. It reported that 422,000 robots were shipped across the world in 2018 – up by 6% compared with 2017 shipments.
According to Robotics Business Review, an online site that delivers “actionable business intelligence” to global robotics, artificial intelligence, and unmanned systems ecosystem through market-leading digital media products and live events, the report is expecting an average growth of 12% per year from 2020 to 2022. Of those robots worldwide, 41% are autonomous guided vehicles (AGVs), 39% are inspection and maintenance robots, and 15% are service robots.
Junji Tsuda, president of the IFR, stated that they saw dynamic performance in 2018 with a new sales record despite some difficulties in the automotive and electrical-electronics industry. “The IFR’s longer-term outlook shows that the ongoing automation trend and continued technical improvements will result in double-digit growth – with an estimate of about 584,000 units in 2022,” Tsuda said.
Helping Robots Navigate Like Humans
One of the abilities that researchers have taught robots is to find their way in a particular environment. In 2018, MIT researchers developed a new navigation system, allowing robots to navigate their environment like human beings. Humans usually go from one place to another safely without thinking too much. They can learn from other people’s behavior and avoid any obstacles. Robots, on the other hand, struggle with such navigational concepts.
In a study titled “Deep sequential models for sampling-based planning,” the MIT researchers demonstrated the advantages of their model in two settings: how to navigate a room with traps and narrow passages and how to navigate an area while avoiding collisions. The new system allows robots to explore their environment, observe other agents, and apply past lessons learned from previous trials to their current object.
“Just like when playing chess, these decisions branch out until [the robots] find a good way to navigate. But unlike chess players, [the robots] explore what the future looks like without learning much about their environment and other agents. The thousandth time they go through the same crowd is as complicated as the first time. They’re always exploring, rarely observing, and never using what’s happened in the past,” co-author Andrei Barbu, a researcher at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), said.
|One of the abilities that researchers have taught robots is to find their way in a particular environment. In 2018, MIT researchers developed a new navigation system, allowing robots to navigate their environment like human beings / Photo by: dolgachov via 123RF|
According to MIT News, the model was developed through a planning algorithm and a neural network that learns to recognize paths leading to the best outcome. It was programmed to navigate different types of environments and not just where they are designed to go. The neural network helps robots interpret their current environment, including physical conditions like the walls and the actions of others in the environment.
The researchers also addressed how robots can handle other agents around them. During these situations, they need to navigate around agents, avoid collisions, and reach the goal (such as an exit). “Situations like roundabouts are hard because they require reasoning about how others will respond to your actions, how you will then respond to theirs, what they will do next, and so on. You eventually discover your first action was wrong because later on, it will lead to a likely accident. This problem gets exponentially worse the more cars you have to contend with,” Barbu explained.
Navigating Without Maps
Usually, AI-powered robots and other devices rely on uploaded maps and precise instructions to get around. This isn’t new because this is how automated factories work. The downside to this is that maps can become outdated. For instance, shifting furniture around would essentially result in the older map being obsolete. To improve this, an AI technology needs to be able to build an internal map of a room using onboard sensors, such as a compass, a camera, and a GPS.
Recently, Facebook introduced a new AI that allows robots to find the shortest route in unfamiliar environments, opening the door to robots that can work inside homes and offices. According to Tech Xplore, an online site that covers the latest engineering, electronics, and technology advances, the Facebook AI created a large-scale distributed reinforcement learning algorithm called decentralized distributed proximal policy optimization or DD-PPO. The tech company stated that the algorithm has effectively solved the task of point-goal navigation using only an RGB-D camera, GPS, and compass data.
|Recently, Facebook introduced a new AI that allows robots to find the shortest route in unfamiliar environments, opening the door to robots that can work inside homes and offices / Photo by: Anna Om via 123RF|
Aside from DD-PPO, Facebook AI also used its open-source AI Habitat platform for its "state-of-the-art speed and fidelity,” which serves as a simulation platform to train embodied agents such as virtual robots in photo-realistic 3-D environments. AI Habitat platform is part of Facebook AI’s “ongoing effort to create systems that are less reliant on large annotated data sets used for supervised training."
"Navigation is essential for creating AI agents and assistants that help people in the physical world, from robots that can retrieve an object from a desk upstairs, to systems that help people with visual impairments, to AI-powered assistants that present relevant information to people wearing augmented reality glasses," the researchers said.
Facebook believes that it will be capable of creating robots that can navigate an area without the need for maps as well as reach its destination 99.9% of the time and with only a 3% deviation from the ideal path using the new algorithm. In theory, a person could place their robots in a certain place without a map and it should be able to find its way to its destination.