Autonomous Decision Making Process TRIZ Cases

Neural activations from real scenes guide simulation parameter tuning, cutting photorealism costs while preserving effective AI training.

Precomputed end-of-use time and position data helps match riders to autonomous vehicles faster, improving occupancy without heavy real-time processing.

Third-party communications and sensor analysis help identify autonomous vehicle status and trigger controlled stops or recovery actions.

Fusing depth and thermal sensor data improves obstacle detection and scene classification for autonomous vehicles in rain or fog.

A common wireless control layer simplifies ROS-level programming and coordination across aerial, marine, and land unmanned systems.

Dynamic route planning uses bale-location and field data to avoid existing bales, cut fuel use, and improve pickup efficiency.

Physical product samples are brought by autonomous vehicle to the customer site, enabling in-person inspection without store travel or salesperson visits.

Optical tags and onboard scanning let industrial vehicles detect warehouse traffic rules, avoid collisions, and navigate more accurately.

Multi-frame CNN processing helps autonomous vehicles distinguish brake, turn, and emergency taillight signals across varied vehicle types.

A gondola transfer lets autonomous park vehicles cross roadways and terrain barriers while enabling guided boarding and exit between separated areas.

A double-row tag matrix improves industrial vehicle positioning and directional detection, enabling accurate speed and height control in facilities.

Voice instructions let the vehicle identify, dock with, and transport targets while adapting routes with sensor-based navigation.

A machine learning model infers missing LiDAR intensity from range data and object labels, improving simulated sensor data for training.

Selective ML image processing confirms treatment actions near agricultural targets in real time, improving detection accuracy with less training data.

Multiple ML input schemes detect treatment actions near weeds in real time, improving selective spraying while reducing processing load.

By detecting backlighting in camera images, the machine can stop or adjust its path to maintain accurate external world recognition.

RGBD sensing, semantic segmentation, and trajectory prediction build local grid maps that help mobile robots avoid dynamic indoor obstacles in real time.

Real-time sensor data and a zeroing neural network improve UAV stability by solving motor power allocation faster and more robustly than PID control.

Autonomous control guides freight vehicles through highway port checkpoints and loading zones to cut driver fatigue, cost, and security risk.

When state variance crosses a threshold, the agent requests human action input, improving safety without slowing routine RL decisions.

Distance-based memory remapping lets aisle vehicles bypass malfunctioning sequenced RFID tags and keep position-linked control accurate.

Sensors in mobile objects capture biometric data with subject ID and time, enabling frequent population-level health tracking.

Electronic pedal-to-wheel torque mapping replaces mechanical gears in a chainless e-bike, adapting cadence, speed, and incline for natural riding.

Virtual 3D vehicle models auto-generate labeled multi-view training data, cutting manual annotation time for autonomous detection models.

Control signals from emergency vehicles, barriers, and road signs guide autonomous vehicles through atypical traffic safely and in coordination.

Combining radar range data with camera object cues resolves 3D distance ambiguity and improves environmental models for ADAS and autonomous driving.

Broadcast correction and post-broadcast integrity data to cut GNSS convergence time while protecting precise positioning in safety-critical use.

Sensor-based region sizing and orientation let an autonomous mobile robot target dirty areas without manual positioning or fixed cleaning paths.

Weather and location data are used to predict grass growth and set mowing time, reducing unnecessary cutting, energy use, and sensor complexity.

On-board mission checks reject corrupt or unauthorized commands before and during execution, helping autonomous vehicles maintain control.

Multi-sensor operating-state judgment stores mower position only at true boundary conditions, improving map accuracy and edge mowing precision.

Online learning updates earthmoving vehicle control from sensor feedback, adapting to changing soil conditions and vehicle behavior.

Satellite and differential positioning are checked against a stored work-area map to alarm on mower location faults and support safer navigation.

A server compares delivery counts across line parts to dispatch less-used robot bodies first, balancing battery aging and wear.

An autonomous companion vehicle follows users or vehicles to deliver, collect, and return outdoor equipment without manual transport or storage.

Fuzzy foothold selection uses sensor feedback and path constraints to help biped robots avoid obstacles while maintaining balance.

Removable batteries cut recharge downtime, while roll-angle sensing stops the blade on tilt for safer autonomous lawn mowing.

Rangefinder and IMU sensing combine SLAM, RNN detection, and Kalman prediction to navigate around moving objects while protecting privacy.

A ride-along platform uses visual perception and a core mediator to monitor aircraft state and alert pilots without invasive avionics upgrades.

Fusing LiDAR and RGB data in a shared latent space improves trajectory prediction across structured scenes, varied speeds, and human-machine interactions.

Test actions and sensor feedback update calibration data, letting one autonomy control unit adapt reliably to different robotic tools.

Precompiled task schedules use dependency and compute-unit analysis to cut runtime overhead and improve real-time execution in autonomous vehicles.

A candidate region based on vehicle position and orientation lets an autonomous work vehicle start on a suitable work route without preset routing.

Pre-collected waypoint data lets an autonomous rescue vehicle relay distress signals and reach rescuers where GPS or cellular coverage is unavailable.

Multiple independently controlled wheel sets use sensor-guided steering to shrink turning radius and help large autonomous vehicles clear tight spaces.

Local sensor-based command normalization compensates for unseen vehicle conditions, cutting bandwidth load and latency in remote teleoperation.

Autonomous agents use LiDAR, cameras, and IMUs to detect people and equipment in excavations and maintain joined clear zones for safer clearance.

Doppler radar detects gusts around a high-altitude relay aircraft so flight direction, speed, altitude, and route can be adjusted to prevent falls.

A mobile terminal supplies sensing, AI processing, and navigation commands so a simple robot can achieve robust perception at lower cost.

A common control stack links vehicle interfaces, telematics, perception, and cloud sync to port autonomous farm functions across mixed machinery.