Security robots and robotic surveillance are becoming more and more popular. But real-world challenges, like the infamous "mall security robot" incident, limit their usage.
Security robots that can patrol, maneuver themselves through busy locations, monitor, inform, and warn people, are attracting more and more attention in the market. Several companies, mostly startups, have come up with autonomous machines that can either augment or replace human guards, especially for monotonous or dangerous work and robotic surveillance. Some of these are promoted as security robots for public protection, to protect people and assets in public places. In fact, some of them were effective even to deal wtih COVID-19 related issues.
While this is an industry that security professionals and investors closely watch
, its development has not been without hiccups. Right from the early days when the Knightscope robot ran over a toddler and hit a fountain in a mall, doubts on the efficacy of using robots in public places have persisted.
A Singapore-based systems integrator with a presence in several countries claims that using a security robot for public protection is a bad idea. What’s significant is that Patrick Lim, who is the Director of Group Strategy at Ademco Security, says this after having attempted to use robots for public safety. He listed the following problems.
“We began using security robots in 2015,” Lim explained. “We separate ground and non-ground (drone) robots for our projects, and we had begun with a ground robot at that time. We experimented with using it till the end of 2016, which was a good full year of testing and trials on the ground. We used several different products from different companies to understand which would suit us best.”
The critical thing to note is that Ademco did not have an issue with navigation, tracking, or any other technology features. In fact, the real problem was not technological at all.
“The problems were more practical,” Lim continued. “In many cases, the terrain was an issue. The terrain that we may want to deploy a robot in may not always be ideal. It is tough for customers to change the terrain just to deploy a robot. So, in places that have issues like uneven ground, we had to change the track for the robot to work. Luckily the main robot that we tested out then was able to function in the uneven ground. Still, we realized that some of the others were too tall and narrow at the bottom that they toppled over easily on uneven ground.”
Ademco had trialed robots of varying sizes and weights. The smallest they tried was about 12KG heavy and was quite small. The largest weighed about 180 KG. The problem with the former was that people could easily pick it up, take it away or toss it aside. The problem with the latter was that its large size limited flexibility to navigate through crowded or narrow areas.
“The large robot used to get stuck in crowded areas or get in the way of people,” Lim said. “But perhaps an even bigger problem is that people, when they see a robot, like to gather around it out of curiosity. And during night patrol, we would have some people who would throw rocks at it or block its way, until there was a voice from the robot that warned them that they could be seen.”
A voice would help the robot get away from one bunch of miscreants, but it would soon come across another. This meant that the robot, which was meant to perform one round of patrolling in two hours, would end up doing it in two and a half hours.
Flexibility of purpose
A third major problem that Ademco encountered was that many of the robots came designed for fixed applications with a narrow focus. For instance, a security robot in a parking lot may have features like a number plate recognition system to deal with vehicles. But it would only be useful in that specific area. If you wanted to extend the robot’s functionality even slightly out of its comfort zone, it wouldn’t be possible.
“You may be able to use a small robot in a mall, hoping that there wouldn’t be any change in the layout of the mall,” Lim said. “But this is not practical. Layouts are bound to happen at some point, and the robot would struggle then.”
The solution that was reached
Finally, after having tried and tested different robots in various scenarios, Ademco decided to move away from ground robots to autonomous drones. Lim explained that although there are regulatory hurdles when it comes to drones, they don’t have any issue dealing with the terrain and are often too high in the air for people to reach.
“Of course, there are some concerns to take care of,” Lim explained. “For instance, what should be done if a drone stops functioning mid-air? There should be a parachute that gets automatically deployed to make sure the machine lands safely. So, I wouldn’t say that drones are completely devoid of any issues, but with them, we were able to overcome a lot of the problems that we faced with robots.”
Lim’s perspective throws light on practical issues that customers would face when using robots for security. Too often, the technological features of a robot get prominence. But what gets forgotten is that these features would work well only in the conditions that manufacturers design them to. Unfortunately, the real world could often be unpredictable and difficult for a machine to be prepared.