In artillery, serving a piece at its rate of fire is learned in the field, on a real system, through repetition. The problem is not pedagogical. It is economic and operational: real firing slots are rare and expensive, the shell weighing over 40 kg does not forgive a mistake in gesture, and each system taken out for training leaves operational service. To prepare its crews without spending ammunition or shortening the lifespan of the tube, a major European land defense industrialist has developed, with Myxed, a mixed reality simulator: a physical replica of the back of the cannon, a 3D scene overlaid on the real world, a headset connected to a secure computer. From a wooden and PVC tube prototype presented in 2018, the device has become a product sold for export, with the first delivery in 2025. A crew fires about one hundred real shells per year. In simulation, they fire several thousand in the same period.
A large-scale training challenge
A piece crew, with up to five positions depending on the versions, operates the 155 mm self-propelled gun mounted on a truck. Supplying, aiming, firing: each position has a role, in a precise order, at a rate that can reach six rounds per minute. The ammunition weighs over 40 kg. The action is heavy, timed, collective. It resists video supports and manuals, because it is acquired through the body and the coordination of the group, not through reading.
This industrialist produces and delivers this cannon all over the world, at an increasing pace, with orders rising sharply in Eastern Europe. Each delivered system is a new crew to train, quickly, if possible before the equipment arrives. These systems remain in a tight flow and in operational condition maintenance. Mobilizing a cannon for training means taking it out of service.
Live fire also has a direct cost. The tube has a limited lifespan, counted in number of shots. Ammunition is expensive. Safety zones are strict, the object is pyrotechnic, a handling error has serious consequences. In France, a crew fires only about a hundred shells per year for training.
A final risk relates to the fidelity of the training. Practicing on a less faithful system instills bad reflexes. On the real system, these reflexes are costly, and the training misses its goal. The military simulation doctrine has changed in recent years: in high intensity, mobilizing real equipment as training support is no longer sustainable.
The trigger
The manufacturer presents a first prototype in 2018, at a major international defense exhibition: wooden boards, PVC tubes, virtual reality headsets equipped with stereoscopic cameras for a makeshift passthrough. The feedback is enough to launch the next steps. In 2019, its training subsidiary opens a demonstrator funded from its own resources. Since 2008, this subsidiary has been designing technical content in virtual reality, augmented reality and mixed reality, for military and civilian clients. The team stabilizes the physical part, weighs the ammunition accurately, and replaces the original setup with the first real mixed reality headsets.
The team selects a priority use case: the crew firing scenario. The sequence starts with the cannon in battery. The movement and cabin are set aside, the training focuses on what takes the most instructor time, exposes the most risk on material, and requires the most coordination. Development then lasts several years, with operational units and training teams.
A physical replica, an overlaid 3D scene, a real-time overlay
The device is based on a 1:1 physical replica of the back of the system. It only reproduces the elements that the crew manipulates: mobile sets, work surfaces, ammunition at real weight. The physical replica is useful when the gesture must be exact to the weight and position, which a handheld controller does not allow. In virtual reality alone, the crew member's hand would pass through the piece.
The headset overlays a 3D scene onto this real structure. The crew member touches the metal where they need to act, and sees around them the complete system, the firing position, and the battlefield in three dimensions. The real that they manipulate and the virtual that they observe occupy the same space.
The system calculates the compositing in real time, frame by frame, without a editing phase. The team paints the neutral surfaces of the room and the replica blue, and not green as in cinema: the combat uniform is green, and the green compositing would make it disappear into the background. In the image, the real elements remain, the blue background gives way to the calculated environment. The process requires diffuse and uniform lighting; otherwise, the shadow areas do not composite.
The headset is a professional mixed reality model, with high-resolution video passthrough and front cameras. The computation runs on a secure computer; the headset only connects to it. The whole system operates offline, on production lines compliant with defense standards. A pulley system carries the cable, and the headset fits into a real gunner's helmet: the crew member feels neither the weight nor the discomfort of the equipment.
The whole team trains at the same time. Each team member has their helmet, their station, their marked and non-crossable area on the ground, their role in the procedure. The team members carry out the complete firing sequence and coordinate in a room of about 100 m². The noise level rises to 92 dB to reproduce the sound and vibration of the shot.
The team designed the device around this constraint of the collective. A crew does not progress station by station. They learn to maintain the pace as a group: absorb a supply delay, restart after a breakdown without breaking the rhythm, keep the tempo under the noise. The simulator reproduces this group pressure, session after session, without a single shell leaving the stock.
The scene changes according to the theater: desert, forest, snow, night mode. The fog degrades visual communication between team members, dust rises at the shot's departure. The simulator reproduces the physics of the system: the cannon does not react the same way at twenty degrees below zero as it does at forty degrees of heat. The instructor triggers breakdowns and disrupts the choreography to instill reflexes. A button resets the scene to zero, without risk to the team or the equipment.
What changes for the instructors and the crews
Before this simulator, training at the pace required a rare firing slot, a real system available, and a risk taken at each session. Part of the instruction time was used to reinstall the fundamentals, session after session.
With the device, the crew repeats the firing sequence as many times as necessary, at the real pace and with the real weight: it can train on critical equipment without taking it out of service, without consuming ammunition or wearing out the tube. The instructor no longer replays the basics. They work on reflexes, coordination between stations, and breakdown management.
For the debrief, the instructor relies on the data from the headset. The headset records the position, gestures, sequence, and adherence to timing. The instructor has access to video feedback, the system status, and a complete timeline of the sequence. He reviews each crew member, then the piece as a whole. He checks if everyone was in their safety zone at the moment the shot was fired. Crews can train even before the system is delivered, or maintain their skills between two live firing campaigns.
The change is also organizational. Instructors work with crews that have already gone through the procedure dozens of times. They start from a more advanced level and dedicate their time to what only an instructor can convey. Training a complete piece, at the right pace and with the correct gesture, without mobilizing an operational system or exposing anyone: the device makes it possible. The simulator prepares, the live fire validates.
Measurable results
Under normal conditions, a crew fires about one hundred live shells per year during training. In simulation, they fire several thousand in the same period, at the real pace and weight.
Training focuses on the complete crew — up to five positions — in coordination. The pace depends on the synchronization between positions, and that is what the simulator makes them repeat, position by position and all together.
The 2018 prototype, made of wood and PVC, has become a self-funded product, sold for export, and delivered for the first time in 2025. Outside of this program, the aerospace sector reports that comparable devices, applied to mixed reality pilot training, would have reduced training time by about two-thirds and allowed for better results than a conventional curriculum.
Scaling up
The system is deployed on the client site, application school, or regiment, with the army providing the space and infrastructure. Instructors first train on the system, then train the crews independently, as in any mixed reality training for industry. The model applies to other systems and environments beyond this cannon. The same principle — preparing a precise, repeatable gesture at scale — already applies to transmitting the craft skills of luxury leather goods.
The next levels are under study. Physiological sensors could measure cognitive load and track gaze. A layer of artificial intelligence could generate scenarios from field data, adjust difficulty in real-time according to the learner's state, and preemptively signal risky actions. These avenues remain a roadmap.
The foundation is in place: a specific use case, shooting; a minimal but faithful physical replica; individual and collective tracking of each session. It is designed to expand.
What has been built does not replace real shooting. It prepares for it.