Autonomous Solutions

Most de-escalation training happens on a screen. ADRT is a training robot that stands in the room.

Name: ADRT: Autonomous De-escalation Robotic Trainer in Action
Uploaded: 2026-08-19T19:46:36.365Z
Description: A trainee steps into the room, and the robot talks back, tracks their movement, and reacts to how the conversation actually goes. This is de-escalation training without a script.

A human-like robotic trainer with voice interaction, physical tracking, and scenario-driven behavior, giving officers repeatable, measurable practice against a subject that talks back.

Talk to our team

Autonomous De-escalation Robotic Trainer

Most de-escalation training happens on a screen: a pre-programmed video with a branching menu, no real conversation, and no consequence for getting it wrong. ADRT is a human-like robotic trainer that stands in the room instead, giving officers repeatable, measurable practice against a subject that actually talks back.

See ADRT in Action

A trainee steps into the room, and the robot talks back, tracks their movement, and reacts to how the conversation actually goes. This is de-escalation training without a script.

Filmed during an actual training session, not a staged demo. This is the baseline interaction every ADRT session runs on.

What ADRT Does

Screens can't hold a conversation. Live actors can't scale, and they're never quite consistent from one session to the next. ADRT is built to close both gaps at once.

Voice interaction lets the trainee actually talk to the subject, the one part of de-escalation a video can't replicate

Physical tracking reads body language and movement in real time, not a scripted camera angle

Scenario-driven behavior means the robot reacts to what the trainee does, not a fixed branching menu

Sensors capture dialogue, body language, and physical movement together, so the debrief covers the whole interaction, not just the outcome

Built on artificial intelligence and machine learning trained on known behavioral patterns from real incidents

Scenarios update as the needs of law enforcement and the community change, so the training stays current

A Screen Can Only Play Back. ADRT Talks Back.

Branching video is the industry standard everywhere else: pick a path, watch a clip, hope the trainee clicked the right option. That's still just a menu. ADRT skips it entirely. The trainee has an actual conversation, and the robot's response depends on what they say and do in the moment, not which button they pressed.

Built on Real AI/ML, Not a Script

ADRT's behavior runs on the same AI/ML practice behind everything else SimIS fields, trained on known behavioral patterns from real incidents, not a fixed decision tree. See how SimIS builds AI/ML for more on the practice behind it.

A Simulator Can Render a Subject. It Can't Put One in the Room.

ADRT stands at a real distance from the trainee, and that changes what's being trained: how close to get, where to stand, when to step back, when to bring a second officer forward. Proxemics isn't an option in a scenario menu. It's the judgment the officer is there to practice, and it only exists when there's something physically present to approach.

Because ADRT is in the room, everyone in it sees the same subject with their own eyes. Contact and cover work the way they do on the street, an instructor can stop the scenario and point at what just happened, and no one has to put on a headset. The caster base moves ADRT into the rooms an agency already trains in, so scenario-based training runs where the work happens instead of in a dedicated simulator bay.

And a robot that runs four hours on a charge doesn't get tired, doesn't improvise, and doesn't have to be scheduled. The same scenario runs the same way for the twentieth trainee as it did for the first, which is the one thing you can't ask a live role player to do.

ADRT is one of several platforms in our Autonomous Solutions practice.

Specifications

Subject	Realistic mannequin, interchangeable across ethnicities, genders and clothing.
Face	An integrated display renders the face, its expressions and lip-sync.
Speech	Integrated speakers simulate speech.
Dialogue	A microphone and dialogue system carry a conversation with the trainee.
Gesture	Mechanical arms with gesture capability.
Behavior	Robotic software controls motion, and artificial intelligence generates the human-like behaviors.
Sensing	Lidar senses the surroundings, so the robot responds to the officer with both voice and movement.
Movement	Robotic sensors for advanced movement.
Runtime	Integrated batteries and charging for four hours of operation.
Mobility	Caster wheel base for moving indoors.
Control	Mobile app for scenario selection, remote control and feedback.
Scenarios	Updated on an ongoing basis as law enforcement needs change.
Status	Patent pending.

Integration details are in the brochure below, built for the training officers and program managers who need the full picture before they need the demo.

Download the Brochure

An ADRT de-escalation training robot with a screen-rendered face, standing in a room facing a law enforcement officer mid-conversation.

What is a de-escalation training robot?

A de-escalation training robot is a physical, human-like trainer an officer practices with in the room, rather than a scenario played back on a screen. ADRT responds with voice interaction, physical tracking, and scenario-driven behavior.

How is ADRT different from screen-based training?

A screen can only play back a fixed scenario. ADRT talks back, reacting to what the officer says and does, so the same scenario can go differently each time.

Does ADRT use scripted responses?

No. ADRT is built on real AI and machine learning rather than a fixed script, so practice does not become memorization of one correct path.

Who is ADRT for?

Law enforcement and other agencies that need repeatable, measurable practice in de-escalation and verbal conflict resolution.

Where does ADRT fit in a training program?

De-escalation sits at the near end of the use-of-force continuum, before any force option is on the table. That's the part of an encounter ADRT trains: the approach, the conversation, and the decisions an officer makes while both are still open. It's built for realistic, interactive and replicable practice, and the mobile app selects the scenario, controls the robot and captures feedback for the debrief.

Is ADRT for academy training?

ADRT is built for law enforcement preparing to enter the field, which makes academy and pre-service training its primary fit. The mannequin is interchangeable across ethnicities, genders and clothing, so scenarios can reflect a range of subjects rather than one default, and scenarios are updated on an ongoing basis as law enforcement needs change.

Reach Out

Want to learn more about how ADRT can serve your training requirement? Ask about specifications, pricing, deployment timelines, or how ADRT fits alongside the rest of our Autonomous Solutions practice, and we'll respond quickly.