“You’re the first officer on the scene. You get teleported in and you’re at an office building. You’re moving down the hallway and there’s an avatar on the ground who’s been shot and they’re saying, ‘Help me, please help me.’ At that point you, the user, have to make a decision.”
So explains Oliver Noteware, co-founder and CEO of a New York-based startup called Street Smarts VR. Street Smarts makes virtual reality training technology. For police officers.
The idea of virtual reality training systems is, essentially, the same regardless of whether it’s surgery or working in a customer services call center: The user dons a VR headset and is placed into a virtual environment, reconstructing a typical scenario that they might deal with. They then get to practice it repeatedly, gradually building up confidence, experience, and, eventually, expertise.
Street Smarts’ technology does the same thing for law enforcement. The difference is in the range of scenarios on offer in its immersive 360-degree training environments. Also, the fact that — as almost goes without saying — police training has been put under a well-deserved magnifying glass here in 2020.
The idea of being able to place a trainee officer into a particular scenario and safely examine how they react; then play it over and over again to ensure that they get it right (or find out if they won’t) is emotionally charged in a way that few other training scenarios could be.
“By recreating the physical, emotional, and psychological stressors and scenarios that they might find themselves in, they start to develop a comfort level in that situation,” Noteware told Digital Trends.
It’s still early days for the VR training system, but the signs are good when it comes to adoption. Its virtual reality scenarios have already been put through their paces by the New York Police Department and Louisiana Police Department. The Texas Municipal Police Association, which has approximately 30,000 members, took delivery of the VR system this week. And the company has just received a “seven-figure contract” from the Department of Defense to further develop its tech.
The importance of proper training
Problems like police brutality won’t always be down to poor training, of course. They could be due to the presence of individuals who should never have been allowed to put on a badge and uniform in the first place. Or it could be a systemic problem in which a culture is perpetuated where such behavior is tolerated, normalized, or even celebrated. But poor training certainly doesn’t help.
As Noteware explained, over the past few decades the number of incidents police are expected to respond to has greatly increased. This isn’t simply about the quantity of said incidents, but the range of incidents. There are domestic abuse incidents, terrorist incidents, overdoses, active shooters, and plenty more. In each of these, one wrong decision can be fatal.
Unfortunately, training has not kept pace with the growing number of jobs police are expected to deal with. Instead, budgets and training have been slashed, leading to smaller, under-funded departments with high levels of turnover.
Noteware is well aware of what substandard training can do. Before he received his MBA at Columbia Business School, prior to starting Street Smarts with co-founder Alice Formwalt, Noteware was an infantry officer in the United States Marine Corps.
“Before we deployed, a lot of the training that we were given on the simulator systems that we used was pretty rudimentary,” he told Digital Trends. “They were flat screens. You point your gun and just pull the trigger. It creates training scars because it conditions you to just look straight ahead; don’t look to your sides because there was nothing there. [It also trained you] to just shoot because you knew your instructor was behind you with a sheet of paper just waiting to see if you shot at the right time. There were no performance measurements. There were no statistics. It was very, very basic.”
This wasn’t a case of no training, Noteware said. In fact, it may have been even worse than that. It was negative training. These low-tech, but still costly, simulation systems encouraged people to always shoot, to not be aware of their surroundings, and to stand still (because the simulation didn’t allow them to move freely.)
When he partnered with Formwalt, a former firefighter and EMT who had designed pilot simulations for the Air Force, they both agreed that VR could be a valuable tool for training. Law enforcement seemed a logical area to aim for. “[At that time] in 2014 and 2015, there was a wave — not quite as powerful as in the last two months here in the U.S. — but there was a lot of attention paid to the issue of police use of force,” Noteware said.
Branching scenarios for every occasion
To support the VR police training scenarios it has developed, Street Smarts VR has created replica police accessories such as flashlights, batons, guns, tasers, and pepper spray, which can be used in the virtual world. “They’re visible throughout the entire scenario,” Noteware said. “When you look down, you’re going to see your tools right there on your belt.”
The scenarios the company creates range from routine traffic stops to active shooter situations to suspected domestic violence call-outs. Each scenario involves a branching narrative with up to fifteen possible branches. These branches are not time-based or linear, but rather are controlled by the instructor, who can skip backward or forward as they desire. Not only does this allow the instructor to vary the training scenario, but also to exert more realistic control over what happens. Realism is vital, but it’s something that an A.I. simulation simply may not be able to match.
For example, a situation involving a victim or potential suspect could play out very differently if, instead of paying proper attention to the person the police are interacting with, the officer is instead looking around the scene in a way that might appear dismissive. Or what if, while they may be acting in an exemplary manner in every other way, they keep their gun or pepper spray drawn the whole time? These are the kinds of details a human instructor can watch out for — and reflect in the way that the situation unfolds.
“Our goal is to create training scenarios that mimic the real world to create that realism and stress and [sense of], ‘oh, man, you have to make a decision,’” Noteware said. Where some training systems focus on the use of weaponry in all scenarios, Noteware said that, with Street Smarts’ scenarios, “There are always opportunities to de-escalate [a situation].”
After playing through a scenario, participants then get the chance to review it as part of a debriefing opportunity. This gives instructors and trainees the opportunity to re-watch everything they did, from where their weapons were pointed to the exact words that they used.
“The system allows for a review of the scenario at different angles so the officer will be able to see their actions from different perspectives,” Bryan Flatt, a training coordinator who has used Street Smarts VR’s technology before, told Digital Trends. “This is not possible with a two-dimensional flat screen simulator.”
“I can say that everyone was very impressed with the product,” Hunter Martaindale, director of research at ALERRT, one of the groups trialing the technology, told Digital Trends. “[Street Smarts] developed a realistic active shooter scenario for us based on a 360-degree video recording we provided. We had professional actors play the role of a shooter and multiple victims. SSVR took that footage and recreated the entire environment. We had participants go through the first phase with professional actors. We collected saliva and blood samples before and after their run. We will [now] compare these values to a different set of participants that go through the VR version to hopefully show that VR training can elicit the same physiological response as a [traditional] hyper-realistic scenario.”
In the meantime, Street Smarts is working to further refine its technology. The team is adding further evaluation metrics that will help interpret its virtual scenarios in a more data-driven manner. This won’t be an algorithmically graded process (it’s far too complex and variable for that), but it could shed light on details a human examiner could miss, especially when it’s buried in a chaos of correlations over time. For instance, is a cop jumpier and therefore likely to pull the trigger at night than at day? Are they more or less likely than a colleague to consistently reach for their gun or taser or to try and calm a situation down verbally? All of these data points may be invaluable as part of a cop’s training.
“[It] allows us to create some performance measurement and data analytics to start to really understand, okay, how did [the officer perform] in simulations at night or daytime or with car stops or emotionally disturbed people [and so on,]” Noteware said. “You can aggregate some data and start to really understand what their performance looks like.”
And, hopefully, to help change it for the better.
- Facebook’s new A.I. takes image recognition to a whole new level
- Astronauts are using VR to train for the Boeing Starliner capsule
- Facebook is training an army of malicious bots to research anti-spam methods
- A literal minority report: Examining the algorithmic bias of predictive policing
- The best Oculus Quest games