Los Angeles Police Dept Puts VR Simulator to Work

A trend has developed as the cost of sophisticated virtual reality (VR)-based flight simulators is encouraging aircraft operators to buy their own simulator instead of sending pilots to training centers. The Los Angeles Police Department’s Air Support Division (LAPD ASD) is one of the latest organizations to incorporate its own VR simulator, a Loft Dynamics Airbus AStar/H125, and the benefits are significant.

The LAPD ASD is unique, not only for the large area that the city encompasses but also because it keeps two helicopters in the air for most of each day, patrolling the department’s 21 divisions in three sectors. One sector is the San Fernando Valley, and it is covered by one AStar crew consisting of a police officer pilot and a tactical flight officer (TFO). Hollywood Hills and south of the I-10 freeway are the other two areas. The ASD operates 16 AStars, a former fire department Bell 412 used for tactical insertions, and a Beechcraft King Air turboprop based at Van Nuys Airport mostly used for prisoner transport.

Also unique for any helicopter operation is that the ASD trains helicopter pilots from scratch, taking experienced police officers from the LAPD ranks and starting them right out in the AStar with no previous flight time. The Loft simulator is being used for some of this training and clearly offers benefits for practicing emergency procedures and maneuvers, including sling loading.

With EASA level 3 and FAA level 7 qualifications, pilots training in the Loft H125 simulator can receive loggable credit for some training. In Europe, a private pilot applicant can log five hours in the Loft simulator, while in the U.S. the allowance is two and a half to five hours, depending on the type of training program. For U.S. instrument rating training, loggable time allowances range from 10 to 30 hours, depending on the qualification of the simulator and the training program. In Europe, the Loft simulator can be used to log some time loggable towards the instrument rating, and proficiency checks and type training can be done in the device.

The ASD 412 is the only IFR-equipped helicopter in the LAPD fleet. Some of the AStars and the 412 are night vision goggle-capable, and ASD technicians do the 180-day inspections of the unit’s goggles, which saves time sending them out for mandatory maintenance. 

All ASD pilots receive quarterly checkrides from one of the unit’s eight check pilots, who also conduct post-maintenance test flights. A full-time safety officer manages the unit’s safety management system. Incident reports go to the safety officer and then get addressed during a safety committee meeting, which may result in a recommendation that will be shared with the ASD pilots during the daily roll call. Every mission starts with a flight risk assessment.

Full-down Autos

While LAPD pilots are allowed to perform full-down autorotations in the helicopter and still do so, the Loft simulator has added a new dimension to autorotation training. Full-down autos in a helicopter must be done at an airport but in the simulator, the instructor can fail the virtual engine in a much worse environment—for instance, during a mission while at a call for service. This adds a much greater sense of reality to engine failure emergencies.

There is an ongoing debate about the benefit of motion in simulators, but ASD chief pilot Kevin Gallagher is firmly on the side of the Loft’s six-degrees-of-freedom motion base after using the simulator since it was installed in October 2024. “The real value is IIMC [inadvertent flight into instrument meteorological conditions],” he explained. Motion may not be as critical in fixed-wing simulators because airplanes are inherently stable and it is much easier to remain in control when losing outside visual references because the airplane continues flying in the same trimmed attitude. A helicopter is inherently unstable and, if not equipped with a stability augmentation system or autopilot, quickly diverges from whatever attitude it is in if the pilot is not paying rabid attention. This becomes vastly more difficult in IMC. “This simulator, if it wasn’t full motion, would be useless,” he said. “Instability is what it’s about.”

The ASD has developed a “surviving IIMC” curriculum and trains pilots on emergency procedures every quarter, including in-helicopter emergencies all the way to the ground, where safe to do so.

Before acquiring the simulator, ASD pilots would also train annually in full-motion simulators at FlightSafety International in Denver. “That’s two pilots for five days,” Gallagher said. Having the simulator enables not only training for the 30% of emergency scenarios that are too dangerous to practice in a real helicopter but also a higher frequency of training because the simulator is onsite at the ASD facility. “Training has to be realistic and frequent,” he added.

The LAPD pilots find the Loft simulator better for depth perception and being able to see height above the ground, which is inherent in Loft’s VR design. “The technology has come a long way,” Gallagher said.

One example of how the simulator benefits training is showing new pilots why they must orbit a scene instead of taking on additional risk by trying to hover out of ground effect at 400 feet. Losing an engine in that configuration doesn’t offer a lot of opportunity to recover. “We can expose them to things [like that],” he said.

Another advantage of VR simulation is that upgrades can be incorporated quickly and Loft’s engineers can customize the simulator to replicate the customer’s helicopter configuration. For example, the LAPD AStars are fitted with Tyler Special Operations Platforms for carrying SWAT teams for rapid insertion. Loft Dynamics is adding VR platforms to the simulator so the helicopter looks more like what the officers are flying.

Loft is also adding the ASD heliport into the simulation, further enhancing training reality. The ASD heliport has its own fuel farm, which is used to fill two fuel trucks that then fill the helicopters. Two foam turrets are mounted where they can put out fires on the heliport.

Flying the Loft Simulator

During my visit to the ASD, Gallagher ran me through two practice sessions in the Loft simulator, for a total of about three hours flying. I’d flown the simulator briefly two times previously at events where Loft was exhibiting, but this would be an opportunity to evaluate its fidelity for practicing procedures that can’t be done in a real helicopter and trying maneuvers that require focused attention. While these can be done in a helicopter relatively safely, for example, hover autos, run-on landings, and slope landings, they are more comfortable in the simulator because I wouldn’t have to worry about damaging the helicopter. Some maneuvers we’d try simply can’t be done in a real helicopter.

The ASD Loft H125 simulator is equipped with a virtual Garmin G500H TXi touchscreen display, GTN 650 com/GPS navigator, and a second radio. Although I didn’t need to interact much with the avionics during our training sessions, it was neat to see my virtual hand reaching out to touch the controls in the cockpit. 

We started with some familiarization flying, and I lifted into a hover, then climbed out for a traffic pattern and normal approach but with a run-on landing, which was fairly straightforward. The motion base accurately replicated the feel of scraping the skids along the runway.

I did another takeoff and left-hand pattern with an approach to a hover above the runway, then a third with a landing on the airport’s grass runway. We were using Bern Airport in Switzerland for these exercises.

Gallagher had me air taxi over to a slippery concrete surface where I tried to keep the helicopter from sliding after touching down. This is something that would be almost impossible to practice in a real helicopter without a much higher level of risk and would require access to an icy surface. This was also a good demonstration of features that can easily be added within the simulated environment for practicing challenging maneuvers. Keeping the helicopter from sliding all over the place and off the surface was difficult, but it was a fun exercise.

The next maneuver was a departure from the airport and a flight to a hospital helipad in Bern. Landing there was much less stressful in the VR world than the first time I landed at the Downtown Heliport on a building rooftop in Portland, Oregon, in a real helicopter.

From the hospital, I lifted off, and Gallagher directed me to the crash-site scene, where I practiced orbiting at 500 feet.

From there, a short flight took me to the slope park, where various mounds are available with gentle to steep slopes, all marked with the degree of slope. I picked the seven-degree slope and managed not to embarrass myself with a decent right-skid-first touchdown.

An important maneuver to practice and that can be done much more safely in the simulator is the steep approach. I returned to the airport and flew two approaches but descended too rapidly on the first try. A high rate of descent in a helicopter approach can easily lead to settling with power, which is difficult to escape from at low altitude.

Taking off the Varjo VR headset after an hour and a half of flying felt a little discombobulating, kind of like entering a movie theater during daylight and exiting after sunset. The VR world felt so real, and the simulated helicopter is extremely well-modeled and flies with great precision.

Emergencies and IIMC

After a break for lunch, I climbed back into the simulator for some emergency scenarios. This would be a good test of the simulator’s capability to replicate scenarios that are dangerous or can’t be accomplished in the real helicopter.

First up was hydraulic failure in hover. The AS350B2s that the ASD flies don’t have the later models’ dual hydraulic system, so that is what the Loft simulator replicates. If hydraulic power is lost, landing the helicopter is difficult, and so is hovering. With hydraulic failure, the helicopter’s hydraulic accumulators can hold pressure for up to around five minutes. In this scenario, I would lose hydraulic power during a hover and I would have to get on the ground quickly before the accumulators lost pressure. 

Obviously, shutting off hydraulic power isn’t done in the real helicopter. In this case, after the failure, all I had to do was land, and as I did so promptly, I avoided any control issues. 

Hover autorotations were next, and this is a maneuver that is practiced in the real helicopter but does carry some risk of a hard touchdown or worse if done badly. Gallagher reminded me to use the “pedal, pause, pull” technique and to look at the horizon, not in front of the helicopter.

I did two decent hover autos where I cut the power, stepped on the left pedal to keep pointed straight, waited a moment, then pulled collective to cushion the touchdown. Then I did one where he dialed in an engine failure, which was a great experience and, again, completely unsafe in a real helicopter. During my helicopter training, hover autos were always a little nerve-wracking, but I could do them all day in the Loft simulator.

Another maneuver that adds tension to training is full-down autorotations that simulate an engine failure from altitude. Gallagher positioned me at California’s Santa Monica Airport and had me autorotate to the ground from 1,000 feet above the runway. I had never before done a touchdown autorotation, other than observing a demonstration in a real helicopter.

During the first autorotation, I was chasing rotor rpm too much with the collective and couldn’t get it to settle at the optimum speed. This one clearly wasn’t going to work out because I lost focus when it came time to flare and touchdown. On the next one, I paid attention to Gallagher’s instruction to pull a little collective (half to three-quarters of an inch) and just leave it alone while managing airspeed with the cyclic. The rotor rpm settled nicely, although I still had to make tiny corrections with the collective, but I was able to spend more time flaring properly. We set the radar altimeter to call out minimums at the 100-foot flare altitude, and I pulled the nose up into the flare and then pushed forward to try to meet the ground smoothly with a run-on landing. I felt much more confident after achieving three good touchdown autorotations following the first bad one. The next series of maneuvers would prove more humbling.

IIMC Encounter

Gallagher next ran me through some typical standard rate turns, climbs, descents, climbing turns and descents, then 45-degree steep turns. The Loft simulator’s handling felt crisp and tight but fairly light on the controls thanks to the hydraulic boosting. I did a set of these maneuvers first in VMC and then in IMC.

It’s important to highlight a major benefit of the VR simulation environment: the ability to impose adverse weather conditions either gradually or instantly on the simulated outside world. There is no need to wear a view-limiting device, which has questionable value in a real helicopter due to the ability to look below the device through chin windows.

Once in IMC and in the 45-degree bank, Gallagher put me to the test and I failed miserably. Not having flown helicopters in actual IFR conditions, except for training with a view-limiting device, I was not prepared for how a helicopter’s inherent instability makes flying in IMC so difficult.

With far more flight time in fixed-wing aircraft, I had always assumed that a helicopter pilot encountering IMC simply needed to observe the instruments and keep flying by their reference. It’s relatively easy in fixed-wing aircraft, but that is because they are inherently stable. When properly trimmed, an airplane will continue along the same flight path with little input required from the pilot. Transitioning from VMC to IMC can be difficult for inexperienced airplane pilots but it is, I have learned, far easier in an airplane than in a helicopter.

When I had the simulator set up in the 45-degree banked turn in IMC, Gallagher asked me to initiate an IIMC recovery by leveling back to zero bank, adding takeoff power with the collective, and setting a climb attitude with the cyclic. The first time I tried this, I lost control. I overcontrolled while trying to return to a level attitude, pulled the nose too high then pushed it too low, and after a few gyrations was clearly out of control.

Gallagher thankfully stopped the simulation before I crashed the helicopter. He explained that it’s important to focus on each movement and what the instruments are telling me and to move the controls extra carefully. I tried the same scenario again and managed to maintain some semblance of control—it wasn’t pretty—until breaking out on top of the clouds. Flying a helicopter by visual references is so much easier, and I now understand on a much deeper level why autopilots are required for IFR helicopter flying.

To say that this was illuminating is a giant understatement. For years, I had wondered why helicopter pilots had such a hard time with IIMC encounters and kept crashing. Why couldn’t they just look at the instruments and follow their guidance? Now I understand, and in a way that is almost as good as a real IIMC encounter in a real helicopter but with the incredible benefit of not having to try to save the helicopter and my life or, more likely, actually crashing. This demonstration, and the way Gallagher set it up, simply blew my mind. But wait, there’s more.

In the next scenario, Gallagher had me flying left orbits around a scene in downtown Los Angeles. The left-hand orbits are so the TFO, who flies on every mission in the left seat, can see the scene. As we orbited at about 500 feet, the visibility gradually got worse. I had to make a quick decision: either go on the instruments and attempt to climb away in IMC towards, hopefully, better weather or pick a spot and “land and live,” to use the term made popular by the late Helicopter Association International president Matt Zuccaro.

There was plenty of room to land because the scene was next to two wide-open baseball fields, and I chose to land before the fog got any worse. I came in a bit too high but managed to perform a steep approach without an excessive rate of descent and safely landed on one of the fields just as the fog thickened dramatically. This felt like a much better outcome than trying to fly away from the scene in IMC without an autopilot.

For the final scenario, back in clear weather, I took off and flew to Santa Monica Airport. An engine chip light suddenly popped up, and this turned into a good lesson on how to manage the situation to ensure a safe outcome. Instead of flying a normal approach—and we were close enough to the airport that it was the selected spot even if the engine was about to come apart—Gallagher suggested maintaining altitude and then flying an autorotation to the runway. That way, if the engine did fail completely, I would be in a position where I could almost guarantee a safe touchdown. This would not necessarily be the case when the engine fails during a normal approach.

I set up for the autorotation and commenced the maneuver and, indeed, the engine did fail. While I came in a little too fast and used up all the runway, I got the helicopter down safely, thanks to the previous full-down auto practice.

Loft’s Dynamics

Loft’s first full-motion VR simulator was the Robinson R22, followed by the H125. It worked with Airbus to develop a simulator for the H145, and this process should go faster because Airbus has a complete data package for that helicopter. For the R22 and H125, Loft had to conduct flight tests to develop the aerodynamic data and then do more flying with experienced pilots to fine-tune the simulators’ handling and control feel. Loft is also branching into the fixed-wing world with two VR simulators for popular airliners, the Airbus A320 and Boeing 737, according to founder and CEO Fabi Riesen.

“Loft’s goal is not to sell simulators only,” he said. “We want to provide a platform for customers to build faster, safer pilots. On that VR simulator type, you can do all the things you can’t do in real life. It makes it very interesting for experienced pilots. Every single malfunction is in the H125 simulator, and you can combine a lot of malfunctions.”

According to Riesen, there are key elements that make Loft VR simulators better than full-flight simulators, which cost 15 to 20 times as much.

The first is the visual system. “If you have a visual system which needs to represent the real helicopter, you need to have a visual system that, if you move your head, the 3D elements are moving too.”

A simple way to understand this is to look at something like a glass on a table in front of a wall. If you move your head, you can see that the glass moves and the background does as well. In the VR simulator, this behavior is similar, but in full-flight simulators with dome projection systems, the glass and background remain static. “The outside world, outside of the cockpit, it’s just projected,” he said. “It’s simply wrong. If I’m moving the head and it’s not moving, the visual system is wrong. And if the visual system is wrong, I cannot behave like in the real aircraft. I need to adjust to that information on how to fly the simulator.”

The second important element is the 360-degree field of view in any direction. Even a full-flight simulator has a limited visual field—for example, 240 by 80 degrees in a typical helicopter simulator. “You are not having the full immersive 3D, 360 degrees effect,” he said. “[When you [look] around, especially like if you make IIMC, at a certain moment you see the screen in front, and then you know you are in a simulator. With the VR glasses, that's not the case. I think that’s a big, big difference.”

Making the immersion feel real is a challenge for simulator designers, and an obstacle for VR simulation is making the pilot’s interaction with cockpit elements look and feel realistic. Loft developed its Pose Tracking System, which not only lets the pilot see his or her hand and body in the virtual world but also provides haptic feedback when manipulating radio and systems switches, knobs, buttons, and touchscreen displays. The pilot flying in the Loft simulator can view their own body and see shadows play realistically in all respects. At night, an arm will look like a dark shadow, and of course, body parts cover up instruments, just like in real life. Turning on the map light casts even more shadows. “All those things you don't have in the traditional simulators,” Riesen said.

Loft recently added night-vision goggle capability, and this adds an extra element to training because it’s easier to replicate a goggle failure or even failure of one lens.

Vertical reference flying is another benefit of VR simulation because of its ability to project the entire world. Doing so is difficult in a full-flight simulator because the visual system doesn’t encompass areas outside of its visual field. “The screen stops there,” he said. 

Eye tracking is a feature of VR headsets, which use cameras inside and outside the headset to monitor the position of the headset and what the pilot is looking at. Of course, eye tracking is not new in flight simulation but by tracking the pilot’s eye position, using artificial intelligence the Loft simulator can automatically detect the pilot’s scanning pattern, and the instructor can use this information to see where the pilot’s process broke down, for example, in a failed IIMC encounter. 

“If I'm doing sling load, you can see whether they have watched the obstacles,” Riesen explained. “If you make an approach in snow, you know the eyes need to stick on the reference point because the moment the snow starts to blow, you will completely lose everything. You can check if the pilot is really sticking on the reference point until the downwash and whiteout is gone.”

Compared to a full-flight simulator, without the big visual system, the Loft simulator is much lighter, and thus its six degrees of motion platform is much smaller. A traditional full-flight simulator requires a tall building, Riesen explained.

The motion platform is much larger because all that mass is harder to move quickly. “Our motion platform, because it’s smaller, can move up to seven times faster in roll rates and it matches flight data better. I can train close to the ground, such as slope landings. That’s hard to do in a full-flight simulator.”

A new feature for Loft simulators is the LofTWIN demo mode, where an instructor or experienced pilot can record a maneuver that can then be shown to the student, as many times as necessary, to reinforce best practices. “If you have a chief pilot from the OEM who can teach all the pilots how to make the touchdown auto,” he said, “this is the power of the Loft suite.”

Providing access to powerful simulators is not just about saving travel costs, Riesen explained. “We know that increasing the availability and accessibility of simulators makes training more realistic based on the technology but it also makes training more affordable. Loft is running a different business model. We’re not selling the simulator, fire and forget. Our customers aren’t simulator centers. We provide them to the end customers—the aircraft operators and flight schools—and we enable them to make pilots faster and safer. With that change in the business model, the Loft suite becomes a tool for flight schools, operators, and airlines, rather than to have as a mandatory checking tool at a simulator center.”

While some might question the value of tools like the Loft simulator for initial pilot training, Riesen believes that using the simulator produces a safer pilot. Typical private pilot training times average more than 60 hours, and if one-third of that is spent in the simulator, even if all those hours aren’t creditable, they remain beneficial.

“If you have a pilot who is doing everything on the aircraft and a pilot doing the same amount of hours in the simulator, which one is going to be safer? In 60 hours in the real aircraft, you do theoretical malfunctions and no full touchdown autos, no IMC. A pilot with 20 hours in the Loft simulator, that’s a pilot who will experience everything. You can also optimize the training. For an autorotation, you don’t have to fly up again and you can do 20 times more autos than in the real aircraft.”

Loft believes strongly in the effectiveness of its simulator and ran an experiment to see if it would be possible to train a non-pilot only in the R22 simulator. Thirty-three students participated, each logging 10.5 hours and covering everything on the private pilot check ride. A flight examiner then tested each student, successfully, then the students were placed in a real helicopter.

In every case, Riesen reported, “all 33 students had a very linear transfer from the simulator to the helicopter.” They were able to start the engine, lift off, fly a traffic pattern, and even complete a successful autorotation in just 45 minutes of flying time. “It shows you that there is a lot of potential in this,” he said.

The advent of lower-cost (around $1 million) powerful flight simulators not only will benefit flight training organizations but is also leading to groups banding together to buy a simulator instead of sending pilots to the large training providers, Riesen said. Mounting the simulator inside a truck trailer is another option that makes sharing the simulator easier.

In any case, Loft isn’t looking to become a training provider. “We provide them the tools to be more efficient,” he said. “We want to provide an experience as close to the real aircraft as possible. That is the only way to train safely.”