In last month’s issue we looked at some of the many LiDAR companies, technologies and security applications at the Consumer Electronics Show. This issue, it’s time for the best – or at least the most intriguing – of the rest. They fall under the categories of drones, robotics, facial recognition and sensing equipment. (Cyber and IoT developments were too voluminous for this article.)
Drones
If CES is any indication, drones will soon be everywhere. Not only unmanned aerial vehicles, but self-driving cars and trucks, as well as water-borne and undersea contraptions. Bell offered one of the most intriguing displays, presenting a model of a city abuzz with drone taxis that pick up and drop off passengers or deliveries on rooftops. Its Nexus 4EX is a four-duct vehicle that operates either as either electric or hybrid-electric, promising enough room for passengers to work on board.
Various vendors touted longer flight times for the battery-powered drones, beyond the standard 20 minutes. Unicorn of Smart IoT Service (USIS) of South Korea displayed its TB-506A, which lists a flight time of 70 minutes and a payload of 1.5 kg. Morocco-based ATLAN Space was flying the flag for its Core™, a UAV module that replaces human ground control stations with embedded artificial intelligence (AI) “capable of achieving mission objectives through cognitive vision, autonomous flight and contextual behavior when facing uncertainty,” according to the product description. CEO Badr Idrissi claims that its signal can’t be jammed. Out of Hong Kong comes XDynamics’ Evolve 2 aircraft, camera system and ground station combo. Reaching a max speed of 57 mph, the aircraft uses LiDAR sensors to optimize mapping and distance measurement.
Some drone applications surprised CES attendees. Beijing Mobox Technology Co. doesn’t put drones in the palm of your hand, but it gets close. The company introduced the TiMAX, what it promotes as the world’s first AI smart drone watch. The mini-drone that you wear on your wrist is encased in a bulky container, which makes it inconvenient for regular use. Though it’s aimed at recreational use, the drone could conceivably be used for low-profile competitive intelligence by people who are lost and are trying to attract rescuers.
Underwater drones
Referred to as ROVs or remotely operated underwater vehicles – also made a splash at CES. Chasing-Innovation Technology Company demonstrated its Gladius Mini, which resembles a diminutive yellow submarine. A Chasing spokesperson says that the ROV, which can run for up to two hours, is used to perform safety and security inspections on ship hulls and on docks in wharfs and marinas, with the ability to spot contraband.
Whereas Chasing aspired to a submarine look, Robosea of Beijing’s newest ROV mimics the appearance of a fish that other sea creatures won’t want to mess with. Its Robo-Shark looks the part, with motion generated by its tail. Robo-Shark is designed to carry devices such as sensors, monitoring and search equipment and trackers. According to its specs, Robo-Shark can dive to a depth of 2,000 meters below sea level and can remain in salt water continuously for a month and still operate.
AI Edging into the Mainstream
Artificial intelligence on the edge – which greatly speeds data processing and analysis and reduces power consumption – is working its way into many security applications that were showcased at CES. Security consultant Steve Surfaro says that a single Hailo Edge AI processor handled no less than ten streams being processed in real time, including a room entrance, a city center and an airplane takeoff.
“Edge AI processors like Hailo, that can handle diverse processes in constrained environments and at ultra-low power consumption, include AI cameras, drones and facial recognition entries,” says Surfaro.
And vendors are joining forces to integrate AI with traditional security hardware and software. For example, AI chipset vendor Ambarella is teaming with sensor manufacturer ON-Semiconductor and solution provider Lumentum for 3D image processing applications such as facial recognition in security and access control, notes Surfaro. In that case, a new platform leverages a single ON Semiconductor CMOS image sensor to obtain both a visible image for viewing and an infrared image for depth sensing. The Ambarella CV25 AI vision SoC powers depth processing and 3D facial recognition algorithms.
Robotics
Robotics were also prominently on display at CES, but based on the exhibits, advances in the security realm appear to be gradual. Obodroid of Bangkok demonstrated its Smart Security Robot, which conducts automated patrols of sites such as parking and office areas, detects and recognizes faces and objects such as weapons, detects fire and smoke, recognizes license plates and identifies specific human poses, such as falling.
More tailored for security applications is UBTECH’s AIMBOT, which is marketed for securing data centers and power distribution rooms. Its sensor-laden “head” rises like a periscope to capture data up high.
Much more lifelike is the U-Partner U05 Humanoid Service Robot by CANBOT of Beijing. Though not primarily a security robot – it is marketed as essentially a combination interpreter, docent, customer service assistant and butler – the manufacturer claims features that would have significant security applications if the robot achieves what it promises. CANBOT, says U-Partner, recognizes and communicates in about two dozen languages, recognizes faces in microseconds, and, using a blend of sensors, recognizes hundreds of expressions and “perceives the internal and external world like a human being.” If it indeed has that capability, the robot would aid in identifying upset, angry, intoxicated, or aggressive people and trigger intervention before they turn to violence or malfeasance.
Facial recognition/video
Facial recognition software powers robots that can identify emotion, so it’s no surprise that facial recognition providers at CES promise that same capability. For example, Taiwan’s Face Me product can determine emotion, age within five years and gender, according to business development manager Munir Haddad. One of its more intriguing prospects is its use at ATMs to better verify users, which the company is exploring with partners. Face Me also reflects the industry trend of computing on the edge – where processing occurs more quickly – and working across platforms.
Anti-spoofing features prominently in Face Me’s marketing, and the company was just one of many facial recognition firms showcasing this ability. IDLive Face, by IDR&D of New York, for example, promoted itself at CES as “the world’s first truly passive facial liveness technology.” Demonstrations showed the software distinguishing between a photograph and a live person based on a single image of each. “We can take a single selfie, analyze the image, and tell if it’s a real person or a spoof,” says John Amein, VP of Sales. In other words, as opposed to some systems, a user doesn’t need to turn their head, smile, speak, or take some other action to be verified.
Exhibitors at CES also highlighted the multiple uses of facial recognition, and many industry- or task-specific facial recognition firms are gravitating to security applications. For example, Mikara, of Sydney, is a smart-store application that deploys facial recognition, machine learning and other technologies to deliver targeted messaging and displays to customers. As well as letting a retailer tailor service to a particular customer, the application can target known shoplifters and other offenders, says CEO Kuba Tymula.
Similarly, Bangkok’s Lumio 3D Co. uses spherical capture technology to record minute details of the human face for healthcare purposes, such as helping simulate post-operative appearance. Now the company is beginning to explore the security market, according to CEO Borom Tunwattanapong.
With Europe’s General Data Protection Regulation and similar legislation being enshrined into law worldwide, anyone capturing a face on video must take care of how it is used and stored. A range of companies, from LiDAR manufacturers to AI providers, are addressing that issue. For example, Berlin’s Brighter AI has introduced an Identity Protection Suite that identifies faces in video images and overlays a blur or some other pattern. CEO Marian Glaser says that the resulting images can be used for further processing and data sharing without running afoul of privacy regulations.
Detection systems
Some of the most interesting exhibits at CES presented niche detection systems that can identify all classes of material and odors.
Video analytics detects visuals
Siri and Alexa understand audio. Touchscreen technologies detect touch. Taste and scent detection technologies, generally, have been confined to research labs. But at least two companies showcased scent-detection technology at CES: Stratuscent of Quebec and Israel’s Nanoscent.
Technologies that detect chemicals, such as gas chromatography and mass spectrometry, are well established, says Stratuscent CEO David Wu. But these units are large and expensive, or in the case of smaller sensors, detect only single chemicals. Stratuscent’s “electronic nose” is portable and real-time, Wu says, and can detect both simple chemicals and complex scents. “Our sensor has the capability to differentiate between Coke and Pepsi and between one-day-old milk and two-day-old milk,” according to Wu. While food/agriculture, smart homes and air quality control are principal markets, Wu says the product detects volatile explosive gases and can do work fit for neither man nor dog. For instance, he says that Canadian law enforcement is using it to avoid contact with substances such as fentanyl. Other chemicals could kill a dog who takes a whiff, he says.
Israel’s Nanoscent proffered a prodigious proboscis of its own. In essence, it’s bringing the olfactory to the factory. Its initial scent-detection applications focus on safety issues, including industrial scent monitoring on production lines, leak detection, car cabin air quality, and air and water quality.
Taking a more holistic look at substances is Senorics, a startup out of Dresden, Germany. It introduced what it calls a “fully integrated material sensing solution” for both solids and liquids that is about the size of a computer mouse. According to Senorics, it deploys customized organic semiconductor materials that are structured in thin layers to form tiny infrared detectors. Current applications include monitoring lubricants for machine maintenance and identifying macronutrients in animal feed. But CEO Ronny Timmreck says that the company is working on improving the technology so it can determine whether a pharmaceutical is legitimate or fake.
These represent but a few of the exhibitors at CES, but ones with the most intriguing current or potential uses for security. The market will decide which ones prosper and which ones fade into oblivion.