Why it was founded: William Li, a former police officer and Ford Motor Co. executive, had just walked away from his Carbon Motor Co.’s efforts to produce a high-tech police vehicle, when he watched news reports about the school shooting in Newtown. He started wondering what technology could do to put a stop to such events; by April, he’d incorporated Knightscope and begun recruiting a team. In early December, just before the one-year anniversary of the Sandy Hook shooting, the company introduced its prototype.
Available: Now enrolling beta customers
What it is: A semi-autonomous security-monitoring robot
Market niche: A range of security applications, including private security, law enforcement, and others.
Funding: Self-funded to start, but currently raising a $1 million seed round via AngelList
Industry Partnerships: Establishing partnerships with other companies working on sensing technology for the security sector.
RBR’s Take: Knightscope, a one-year-old, eight-person company, arrived on the scene this month with a great deal of fanfare and enthusiasm. The K5 robot, with its imposing height and friendly pear-shaped form factor, has earned comparisons with everything from the cheerful, chirping R2D2 to RoboCop — and even Paul Blart, Mall Cop.
The company’s vision of a “precog” robot for crime deterrence and detection includes everything from radiation and chemical monitoring sensors for air and sea port applications to social media sentiment analysis.
Many of these features, such as 360° cameras and WMD detection, were included on the list of features in the Carbon Motor’s E7 cruisers.
However, the current K5 model is limited to an eye-catching chassis that integrates video capture, optical character recognition, and — as of this week — semi-autonomous movement. The sleek, blue-white-and-silver robot uses all-terrain wheels to traverse outdoor surfaces, and is capable of navigating small curbs.
Essentially, it’s a semi-literate observer capable of wandering a parking lot or corporate campus, reading license plates.
Li says even in this nascent form, the K5’s capabilities are enough to catch the interest of private and public security agencies alike; a dozen entities have agreed to participate in the company’s beta program, which will help drive product development and diversification for different markets.
Private security firms face a number of challenges in both retaining workers and delivering customer results. The industry has turnover rates of 100-400% annually; that’s because the job, he argues, is dull and can’t be done well by a human.
Citing statistics that a human loses the ability to effectively review security camera feeds after 8 minutes, Li suggests that the basic tasks of security monitoring should be left to robots, which excel at precise and tedious tasks.
If the K5 robot is able to collect 90 terabytes of data each year — a reasonable estimate, Li says — that’s more data than a human can process and make sense of. And, with Knightscope renting out its units for roughly $6.25 per hour, the K5 is cheaper than a human.
“What we want to do is to put a complex puzzle together to give someone a better understanding of the situation,” he told me. For example, the K5 could use audio and visual sensors to detect a gunshot, rapid hand movement, and horizontal human figures. When it alerts police to a potential incident, it could also provide a list of suspicious vehicles, number of officers in the area, and other relevant data to provide responding officers with a clearer picture of what they’re about to encounter.
Leaving aside the rich vein of dialogue about the privacy concerns this scenario inspires, Li’s vision is (theoretically at least) possible to realize with existing, off-the-shelf technologies. Autonomous vehicle technology is maturing rapidly, and companies like Palantir are pushing the boundaries of big data’s potential.
While Li says the company is developing intellectual property in both of these areas — low-speed autonomous technology and predictive analytics — the real value the company is bringing to the table is in systems integration.
“People like to say, ‘Hardware is hard,'” Li told me when I visited the company’s offices in Sunnyvale, Calif., earlier this week. “That’s true. If you’ve never done it before.”
Li and other members of the founding team are veterans of the automotive industry. For them, integrating hardware design, materials science, sensors, safety, all-weather design, and more into a single consumer-friendly package is familiar terrain. Carbon Motors may not have succeeded, but “this is three orders of magnitude easier than cars,” Li says.
But while integrating autonomous movement and predictive analytics is a ripe opportunity — the home healthcare market comes to mind, for example — the Orwellian feel of robotics-assisted police forces have a certain “ick” factor that will be hard to surmount. Still, there are plenty of highly monitored work environments that exist today, and the K5 is well positioned to improve the performance of such systems.