Standing at 5 feet tall and weighing 300 pounds, Knightscope K5 looks like it came straight out of the sci-fi film Stars Wars. But the crime-fighting robot is actually a new breed of security guards roaming Silicon Valley, including at Palo Alto’s Stanford Shopping Center as part of a pilot program that ended in June.

Mountain View-based company Knightscope, which was founded in response to the deadly shootings at Sandy Hook Elementary School in Newtown, Connecticut, developed the robots. The shiny white bullet-shaped machines gather real-time data through numerous sensors as well as publicly available data from businesses, government and social data feeds and then process the data through a predictive analytics engine, said Stacy Dean Stephens, vice president of marketing and sales. If the robots determine that there is a concern or threat in the area they are patrolling, a real-time alert is sent to the community and authorities through the Knightscope Security Operations Center.

The robots, which operate within a predefined area, are unarmed but their size and appearance will likely make someone think twice before committing a crime, Stephens said.

“One thing that the robots can do … is provide a commanding presence. Obviously because of their size, they get a lot of attention. If you can imagine a criminal coming onto a property and seeing these robots roaming around. They don’t know what they do, they don’t know what they are capable of,” Stephens, a former police officer, said. “Criminals don’t want to get caught, and the biggest thing you can do to prevent and deter (criminal activity) is provide a commanding presence.”

The robots have many features, including a built-in emergency intercom, proximity sensors, GPS locator, license-plate recognition sensors, thermal-imaging and night-vision cameras and 360-degree video and audio-recording capabilities. They run on lithium ion batteries and return to a charging station to recharge.

The microphones on board the robot allows it to act like a two-way intercom, Stephens said.

“If you have someone in an emergency situation, and they need to contact someone in a security operations center, they can push the button on top of the robot to have a two-way conversation,” he said.

The K5 also has broadcast capabilities, so, for example, in an active-shooter situation, the robot can broadcast a message to warn everyone of what’s going on, Stephens said.

“It can also be something more friendly or benign like telling people that the mall is closing and make your last-minute purchases,” he added.

Stanford Shopping Center, which is owned by Simon Property Group, participated in the test to try out the Autonomous Data Machines. The program started, coincidentally, on May 4, often referred to as “Star Wars Day.” Two robots were deployed at the shopping center to patrol its parking lot and interior space.

Neither a representative of Simon Property Group nor Stephens would disclose how effective the robots were during their trial, nor whether the shopping center plans to continue their use.

But Stephens said the approachable and friendly looking robots (which look similar to Star Wars’ R2-D2) have received an overwhelmingly positive response from people who have encountered them. People will come up and interact with the robots — even give them a hug.

“When you put (the robots) into public areas like the Stanford Shopping Center, we have to make certain that the public embraces it,” he said. “Everyone has been extremely both friendly to the robots but also really enthusiastic about their mission and what they’re there for.”

The robots are deployed at different places, including corporate campuses, shopping and data centers, large sporting venues and big-box stores, Stephens said, but the robots are not meant to replace human security guards. They enhance the way security is done today by assisting officers and improving response times, he said.

“There’s only so much technology an officer can carry on their belt or on their back, so what we wanted to do was augment that in a way that was never done before, and that is put all kinds of sensors and technology on a fully autonomous robot,” Stephens said.

For those concerned about the robots threatening individual privacy, Stephens said the company takes privacy very seriously and the technology the company employs is no different than what people encounter every day.

“It is a security camera. It’s no different than when you walk into a store and there is a security camera mounted up on the door sill. This just happens to be a mobile platform for doing so,” he said.

The real threat is violence, Stephens said, and the K5 is an innovative way to fight crime.

“Most people are unaware that crime has a $1 trillion negative economic impact on the U.S. each year … and we feel like, if we can cut that in half, we’ll be doing something good not only for our local community but also for the entire country,” he said.

Join the Conversation

30 Comments

  1. With all the data these machines collect, there is no mention as to how freely the Shopping Center/Robotics developer will share this data with local police, or the victims of crime at the shopping center.

    Not clear how these machines would be superior to more surveillance cameras.

  2. Exponential technology to serve human needs. Good idea. Should lower the cost of many menial security jobs currently done by humans. If the scanning technology is good enough, then facial recognition patterns might be able to be collected…criminals on our streets could be caught much sooner. Overall, crime should go WAY down.

  3. >A great opportunity for hackers of all shades of gray.

    True, but so is social security and DMV and DOD and our bank accounts and credit cards. The times we live in. Encryption wars will be with us going forward, but that should not stop us from going forward.

  4. I see them at the Knightscope company in Mountain View patrolling in the evenings. Among their other capabilities they can scan license plates and sense fumes.
    I still don’t see the connection to Sandy Hook. Are the deployed at that school now?

  5. Two guys, large blanket. Throw on top of robot. Pick up, put in van. “you want your robot back? Give us $10,000 and we’ll return it.” (it’s gotta be worth more than that)

  6. Anyone stealing one of these robots would be extremely foolish as they are bound to contain gps signal gadgetry and the robbers lair would soon be discovered.

    On second thoughts, perhaps they should steal them so that the police can catch them and charge them.

    On third thoughts, we need these at the bike racks at the schools and Caltrain stations.

  7. Second paragraph: scans social media and has sensors to help predict a threat with its predictive algorithms…

    So what it may do is scan Facebook et al, reads the posts and then if it feels you are a threat looks for the phone of the person, tracks it via gps and uses facial recognition to find them at the mall before they shoot the place up.

    While that is kind of cool in a high-tech wiz bang way, it kind of freaks me out. Imagine if it tagged YOU, based on a random set of posts. Now it is your job to prove you are not a nut. Not an uphill battle I want to wage.

  8. >Imagine if it tagged YOU, based on a random set of posts. Now it is your job to prove you are not a nut

    It could tag me as a nut (others would), but I don’t involve myself in criminal activity. I prefer that the criminals get caught, before they hurt someone, especially my own interests.

    Another use for these robots would be to scan the parking tags in residential parking permit programs, like we have here in College Terrace (and about to start in Downtown). Save a bunch of money.

  9. Funny Doctor Who reference … he does like like a reboot of the Daleks.

    Is there anything to keep them from selling identifications of people they may make in Stanford Shopping Center. That might be very good for companies trying to narrow the net of their advertising, and leave out all the poor people.

    What is this guy going to look like and act like in 20 years?

  10. Recent arrests at Stanford Shopping Center (PAPD log 8/10):
    08/07/2015 15:25 Commercial Burglary, 2 booked at main jail
    08/08/2015 16:46 Credit Card Fraud, 1 booked at main jail

    Robots did not contribute. From what I gather, store employees recognized the suspects from previous encounters.

    In other news, shoplifters may be disappointed to see a security guard now at their favorite yoga/women’s-wear store on University.

  11. @musical:

    The article states that the robots were part of a pilot program that concluded in June. The two reports you list are from this month.

  12. It might be a good idea to have a couple at each high school.

    Then again, they probably cost $500,000 each.

    Then again, considering “Moore’s Law”, in ten years they’ll be cheap!

  13. @Parent, point granted. Just thought I’d give a current example of the problem at hand. Don’t know what the robots could do about this sort of crime, other than try to look intimidating. Might want to program them to recognize that a license plate with paper taped over the numbers is a suspicious circumstance (referring to an allegation that contributed to a couple of the aforementioned arrests).

  14. BTDTGTTS. Apogee Robotics did this. The boss would not spring for faster microprocessors. Too many jobs for the 386/486 Aztech SBC in each robotic we built. Now I could just shove in a Itanium based SBC and fix the lot. Moore’s Law DID work and people have this “ cute “ version.
    You should have better success than we did.

  15. @musical:

    It’s hard to say at what point of development these robots are at, we don’t really know how sophisticated their programming algorithms are. They could be anywhere from mobile security cameras to something more advanced.

    In any case, I don’t expect such a pilot program to exhibit extremely advanced analysis. Right now, scanning license plates and comparing them to a database of stolen vehicles is not much, but it’s better than a human security guard getting paid a little over minimum wage for this task, because the human guard doesn’t have the same tools to cross-check hundreds of vehicles license plates. A robot can.

    In any case, I wish the best of luck for Knightscope. If they don’t make a breakthrough, someone else will eventually.

    There are already highly sophisticated pattern matching algorithms in use, even at a consumer level (face recognition in photo software or online photo sharing sites). How effectively Knightscope has employed these techniques remains to be seen. Heck, if you don’t have a FasTrak toll tag in your vehicle and you drive on the Golden Gate Bridge, a camera will figure out your car.

    The prudent reaction would be to encouraging of such efforts, but also wary of the privacy implications as well as the accuracy and ultimately efficacy of such technology. But the jury is out and it will be years before we can pass judgment on these early efforts.

  16. How far will it roll when tipped? Think of a 300 lb rolling egg!

    It does look like a Dr. Who Dalek with a heavy dose of Apple style, doesn’t it?

    What happens to all those faces and license plates the thing scans and stores all day? Who decides?

    Is it related to the “Digital Egg” in Lytton Plaza?

Leave a comment