A leg up for drivers
Why motion systems make better vehicle simulators
By Alan Dron and Karen Walker
June 01, 2010
June 01, 2010
Motion systems — the legs on which a simulator’s cabin sits and that move the cabin in response to control inputs from the pilot — are traditionally confined to aircraft simulators. Today, however, there is a growing trend to put them on combat-vehicle simulators to provide new levels of realism when training vehicle drivers and crews.
There are snags, however. The type of movement experienced in a ground vehicle differs considerably from that of an aircraft. For one thing, an armored vehicle moving over rough terrain gives a lot more movement cues than an aircraft, which is typically smooth unless it hits air turbulence or is pulling high g-forces. So ensuring that the vehicle driver gets the same accurate feedback, correctly correlated with visual and aural cues, as a pilot gets in a full-flight simulator is technologically challenging. Get it wrong with a vehicle simulator and there’s a real risk that the driver will get motion sickness.
Another problem is that motion systems are not cheap. A high-fidelity, electric-powered six-degrees-of-freedom motion system is a sophisticated device that can cost more than $100,000. That’s viable on a flight simulator that duplicates the controls and movements of a multimillion dollar plane or helicopter, but much harder to justify for a ground vehicle.
Nevertheless, new technology and market forces have led to the emergence of a new generation of tank and armored-vehicle simulators on legs. And just as the flight simulator changed everything in the aviation world, so high-fidelity vehicle simulators are changing how troops prepare for the battlefield. In simulators, pilots practice potentially life-saving skills, such as water landings or engine failure on takeoff, that would be too risky to practice in real aircraft. Similarly, a soldier or Marine can now push a tank simulator to the edge of that vehicle’s envelope in dangerous maneuvers that teach him vital cues that will prevent him from rolling a real vehicle.
For all the similarities on what a motion system brings to flight and vehicle simulators, however, there are some important differences.
An aircraft essentially rotates around its x, y and z axes, and “sometimes you feel a bit of turbulence and that’s it,” said Jack van Hoek, sales manager, motion technology, with Bosch Rexroth in the Netherlands.
“Driving is a completely different ballpark because a car or vehicle behaves completely differently, and the way you drive or manage it is completely different from the way a pilot flies his plane. A ground vehicle is basically making linear movements — back, forward, up, down, left, right — and there’s hardly any rotation apart from body roll.”
A ground vehicle and its driver also experience higher accelerations and more intense frequencies of movement.
“A pilot looks at his instruments and, based on what he sees or feels, makes a decision on what to do. As a driver, you don’t have all those instruments. You feel what the car is doing: for example, skidding off into a ditch. That information is computed in your brain, and you react. So, if you want to make a vehicle simulator, the vehicle subsystems need to be at a sufficiently high level to replicate much more of that feeling.
“For a ground vehicle, you need much higher fidelity motion than you need in a flight simulator.”
David Rees, senior vice president and director of special projects at SAIC, agrees that ground-vehicle simulators place special requirements on the motion system. “You have got to accurately correlate the visual system with all the bumps. And being close to the ground, the driver gets a lot of cues from the physical environment around the vehicle, so accurately simulating that environment is important,” he said.
The U.S. Army’s adoption of the Stryker armored vehicle was an important milestone for vehicle simulators, Rees said. Stryker entered service some eight years ago and was a wheeled vehicle that was fast and agile, but which presented rollover issues. “The issue became how to pre-sensitize people to rollover, and the only way you can do that is with a motion system,” Rees said. Hence, the services began to search for affordable, full-motion vehicle simulators.
Rendering vehicle movements accurately and linking them with the appropriate visual cues is important, said Sunjoo Advani, president, International Development of Technology (IDT) in Breda, Netherlands. If the motion produced by the motion system fails to correlate accurately with the visual representation being projected in front of the trainee, the result can be false cues and the risk of negative training.
Some simulator manufacturers have embodied basic levels of “movement” in their products through such simple devices as seat shakers. However, a full-motion ground-vehicle simulator requires a serious amount of computing power.
“You have to start by knowing the vehicle dynamics,” said Mark Saturno, director of business development at Cubic Corp.’s simulation division.
“You can’t just convert a flight simulator into a ground-vehicle version because it’s all a function of the vehicle dynamics. You’re talking about everything from the vehicle’s suspension to the weight, balance, handling characteristics and span between the wheels or tracks. All that has to be written into the software.”
White noise
A dissenting voice as to the relative complexities of aircraft and ground-vehicle simulators comes from Charles Bartel, market product applications manager of Moog.
“A ground vehicle is easier to simulate in that there’s a lot of acceptable noise — in a military vehicle, you’ve got a big V12 diesel that’s always rumbling away. When you’re in flight, it takes technical expertise to attain that smoothness. It’s easier to simulate a ground vehicle because the movement of a vehicle doesn’t have to be so smooth.”
Among the innovations Moog is introducing to new-generation ground-vehicle simulators is white noise, plus special effects that allow the simulator for a tank or armored vehicle to replicate, for example, the effect of a 6-inch drop that a real vehicle might experience when crossing a shallow ditch. Buffets in the x, y and z axes can be programmed with a certain frequency and amplitude for greater realism.
The most recent major innovation in motion systems was the sudden leap from hydraulic to electrical systems, which reduced operating and maintenance costs of the systems while retaining motion fidelity, Bartel said.
The next few years likely will see a process of gradual refinement rather than any further significant leaps in technology, he said.
MRAP training
A major program, in which Cubic has partnered with Thales of France, is the U.S. Mine Resistant Ambush Protected (MRAP) vehicle category. See sidebar.
“The key thing about our Reconfigurable MRAP Vehicle Trainer (RMVT) is that we can integrate the latest commercial game engine technology. The military is trying to integrate open architecture similar to what you would run in gaming technology. The big leap we made with RMVT is to use commercial gaming engine for the graphics. The challenge is, how do you take high-fidelity graphics but still focus on its training objectives? Some people like the glitz and sexiness of game engines, but you can’t ever forget you have to focus on the objectives,” Saturno said.
The MRAP training objectives for mission-based simulators are still being formulated, Saturno said, but the software being developed for the vehicles’ simulators is pushing the boundaries of complexity.
Mounted on MRAP vehicles such as the Buffalo is a long “interrogator arm,” designed to investigate disturbed ground or grasp suspect packages ahead of the vehicle. The arm is reproduced on the simulator to give the drivers the chance to practice manipulating it.
Frequently, improvised explosive devices are buried. That requires Cubic to perform highly complex calculations in its programming. It has to give the driver the ”feel” —through his controls of the interrogator arm — of digging through a deformed roadway and of resistance as the end of the arm removes virtual “earth.”
“It’s very difficult to replicate that force of the arm hitting the ground and moving the earth,” Saturno said.
The U.S. Army has released draft documentation for an approaching program for a full-motion MRAP trainer; the service plans soon to release a formal request for proposals.
Common driver trainer
At the ITEC 2010 conference in London in May, SAIC featured its Common Driver Trainer (CDT), complete with full-motion system, on its booth in the exhibit floor. The CDT was configured as a Cougar armored vehicle and equipped with a Rockwell Collins EPX visual system. The entire trainer is housed in a trailer that can be driven to where training is needed.
SAIC created a CDT MRAP prototype on its own research-and-development dollars after its success with the Stryker simulator. The CDT’s cab is reconfigurable to a number of vehicle types, such as the M1A1 and M1A2 main battle tank, various MRAP vehicles and the original Stryker. An interchangeable dash panel in the cab provides the correct physical interfaces of each vehicle as well as the software and data unique to each vehicle type. Either deployable in a trailer or fixed at a training camp, the simulators can be networked, and different vehicle types can be mixed and matched to fit the training requirement.
“There are a number of European nations who have identified requirements for this type of trainer,” Rees said.
Russian tanks
CAE, meanwhile, has embarked on a program that led it to build four Russian tank simulators in the hopes of winning a contract from the Indian government. The Canadian company partnered with Indian company Tata to compete for an Indian Army requirement for trainers for its Russian-built T-72 and T-90 tanks. The way India has structured the contest, termed a no-cost, no-commitment competition, bidders were requested to build four simulators for evaluation: one full-motion T-72 simulator, one full-motion T-90 simulator plus a gunner simulator and a command and gunner simulator.
CAE has built and shipped those simulators; its team is competing with another led by Indian company Zen Technologies. CAE officials said they are hopeful a decision might be announced later this year.
CAE’s tank simulators have CAE Medallion 6000 image generators and a six-degrees-of-freedom motion system that includes mechanical parts made by Servo Controls Inc. of India.
“People didn’t used to use motion technology on vehicle simulators because of the technology and cost barriers,” said Marc St. Hilaire, CAE’s vice president, core engineering. And if you are building a tank simulator with a motion system in a marketplace where most companies are offering systems without motion systems, then you have to be able to compete on that basis. For us, however, these systems are our bread and butter. The research and development has been done and we are competitive players because we are able to scale the technology down.”
St. Hilaire said the Indian requirement calls for a motion system, but does not specify a full six-degrees-of-freedom system. It was CAE’s belief, however, that the greater fidelity afforded by its full system will be advantageous.
“We made the decision to go with six degrees because this is the core of our business as a flight-simulator manufacturer so the transition for us was easy. It’s not a challenge from the technological design or production point of view,” St. Hilaire said.
“And going this route enables us to do those special maneuvers on the edge of the vehicle’s envelope, such as driving sideways up and down slopes or bridges where the tracks fall off. This gives a very realistic feeling to the driving.”
CAE created the data it needed to give accurate motion cues through its own mathematic modeling. Accurate and robust modeling of the vehicle’s rough motion cues, St. Hilaire said, was similar to the task of recreating vibration in a helicopter simulator.
A major issue remaining for the growing motion-based vehicle simulator market is that there have never been internationally agreed upon standards for measuring the accuracy of simulator motion systems.
“In new flight simulators, there’s now an international standard that is a way of measuring the performance of the cueing system and comparing it with that of the movements of the real aircraft. Up to now, it’s all been subjective. Previously, all simulators built for, say, a Boeing 737 would all have felt different because an individual pilot would have ‘tuned’ it,” IDT’s Advani said.
“I think the real breakthrough will be when we get a consensus for the criteria for motion cueing. Without that, we’re just pulling numbers out of the air and designing systems that we only think are right.”
“The next step will be to take that standardization into helicopters and driving simulators.”
Van Hoek agrees. “There’s one huge problem, in that there’s no real regulation out there right now for minimum standards of fidelity. For driving simulation, there are not yet any established standards.”
