Autonomous Decision Making Process TRIZ Cases

Real-time sensor analytics adjust steering, braking, and acceleration to help autonomous vehicles handle weather, traffic, and road changes.

Dynamic series-parallel battery switching matches vehicle charging needs while cutting resistor losses and improving charging flexibility.

Infrastructure-guided autonomous parking routes help vehicles avoid collisions, cut parking time, and return to a pickup area without driver search.

Reconstructs road graphs and samples plausible behaviors to generate diverse multi-agent driving trajectories from limited real-world data.

Multi-level sensor ranges let vehicles react to low-confidence distant detections earlier during unprotected turns in fast oncoming traffic.

Dynamic ROI sizing and object trajectory prediction help autonomous vehicles change lanes around blockages without raising collision risk.

An independent monitoring algorithm checks ML control decisions in autonomous operation and triggers a secure state when decision quality is low.

Visual markers, cameras, and LiDAR let an autonomous vehicle deploy a wheelchair ramp, track ingress, and secure the chair without driver assistance.

Tagged timestamps and CPU usage baselines expose latency anomalies, enabling safe-state control and reconfiguration in autonomous systems.

A secondary perception layer classifies corridor objects and occupancy maps to reject unsafe vehicle trajectories with lower compute load.

When a local failure interrupts an autonomous trip, the vehicle selects a safe pullover and transfer area while requesting a second vehicle.

3D cargo sensing and obstacle distance mapping let an AGV adjust its path automatically to keep loaded cargo clear of nearby objects.

Crowdsourced gaze, head, and posture cues improve road-scene behavior prediction beyond motion vectors for safer vehicle response.

When network quality drops, the vehicle stores and uses sensor-derived driving condition data locally for continued operation and later transmission.

Sensor-based distance checks predict when another road user will enter a vehicle path, even when lane maps are inaccurate or missing.

Nearby EVs are matched by location, charge level, route, and timing to supply power to stranded vehicles without fixed chargers.

Map and sensor fusion identifies blind intersecting roads and shifts the vehicle path away from occluded traffic before crossing.

Aggregated human responses train models that predict pedestrian, cyclist, and driver actions in complex road scenes beyond motion vectors.

Reinforcement learning tunes autonomous vehicle speed commands to match human driving patterns, reducing abrupt maneuvers during obstacle avoidance.

Fused radar, LIDAR, and map features improve object trajectory prediction for autonomous driving while balancing accuracy and model complexity.

Interaction-zone scenario augmentation adds realistic object behaviors to focus autonomous vehicle motion planning training on relevant events.

A simple confirmation action lets a follow robot verify the intended person before moving, reducing misidentification and unstable tracking.

Dual brake controllers exchange CAN monitoring data to transfer braking control during faults, improving autonomous vehicle brake reliability.

Hypothetical vehicle points beyond sensor range help autonomous vehicles judge occluded cross traffic and complete crossings safely.

Duplex monitoring pairs onboard obstacle detection with remote camera review to reduce false stops and permit safe restart after interruption.

Obstacle reactions are delayed only when sensor confidence is low and contact distance allows, reducing false interventions without compromising safety.

Multiple command controllers combine object-based speed and steering requests to avoid abrupt maneuvers while improving autonomous driving safety.

Dynamic vehicle coupling forms rigid platoons on shared routes, then separates for diverse destinations to improve capacity use and energy efficiency.

Metric-based driving scenarios rank unsafe AV cases by safe paths, effort, and time to act, improving accident-avoidance testing and training.

A transparent movable deflector shields vehicle sensors from water and debris at speed while preserving light transmission and a clear view.

Mutual importance scoring helps autonomous vehicles focus prediction on high-priority nearby agents, improving planning accuracy under limited compute.

FPGA neural inference offloads camera video processing from the MCU, improving vehicle control capability and reliability for automatic driving.

Environmental cameras track people and moving bodies so robots can update routes in real time, reducing interference and collisions.

A lightweight manned vehicle rides inside an autonomous protective vehicle, using a barrier and coordinated driving to improve crash survivability.

By comparing vehicle route cost with passenger walking effort, the system shifts pickup sides to avoid U-turns and unprotected left turns.

A reconfigurable collision detector updates planning-graph edge risks so autonomous vehicles can optimize paths around uncertain dynamic objects.

Location-specific traffic simulation uses learned driver copies to update autonomous driving policies with more realistic interactions and less road testing.

On-board sensors are fused with V2V output to identify non-connected vehicles and share their position, direction, and speed.

Piecewise linear FMCW chirps let one lidar deliver fine resolution for near objects and coarser resolution at longer range.

Failsafe regions from motion forecasts help autonomous vehicles detect unexpected actor paths faster and adjust movement to reduce collision risk.

Phase-coherent LIDAR and ML estimate nearby vehicle yaw rate and direction, helping autonomous control adjust steering and speed safely.

Automatically generated trajectory labels train one DNN to plan paths across vehicle types, cutting manual labeling and model complexity.

Driver rejection extends the next overtaking suggestion delay, reducing lane-change prompts that conflict with driver intent.

Low-confidence obstacle detections are held only at short contact distance, balancing collision risk reduction with smoother autonomous driving.

By testing SoC memory before command execution, the controller blocks unsafe vehicle actions and triggers safe-mode responses.

Fused object uncertainty heatmaps replace per-object updates, reducing collision prediction cost while preserving drivable-area decisions.

Real-time links between detected objects and planned vehicle actions help occupants understand autonomous driving decisions and reduce unease.

Prebuilt rule models from human driving data help autonomous vehicles handle intersections, stationary vehicles, and jaywalkers with less real-time compute.

A spur-gear steering drive and electric brake actuator cut Drive-By-Wire complexity and space for autonomous vehicle retrofits.

Multiple vehicle observations, heading, speed, and ground truth are combined to correct road sign position offsets and reduce longitudinal errors.

Directional acoustic sensing guides a drone toward unrecognized sound sources, improving low-visibility autonomy with less weight and power.

Propeller ground effect is used as a backup surface sensor, detecting proximity when optical sensors are degraded by rain, fog, snow, or glare.

Parking spots are assigned by dataset size, wireless link quality, and access point capacity to speed autonomous vehicle data transfer.

Real-time wind data from other drones reshapes flight routes to avoid headwinds and use favorable airflow, reducing onboard energy use.

A height-adjustable inverted omega platform helps AGVs load, unload, and assemble parts with compact dimensions and better operator ergonomics.

Coordinated incident response uses occupant, vehicle, and environment assessment plus inter-vehicle data exchange to guide care actions.

Roadside sensors broadcast zone IDs and arrival times for approaching objects, helping vehicles react when onboard views are blocked.

Varying polygon scanner facet fields of view raise sampling density and signal quality for more accurate long-range LIDAR range and velocity detection.

A vehicle computer selects a safe rendezvous point and coordinates simultaneous ambulance arrival to cut emergency response delays.

Pre-processed vehicle state data lets a central platform detect emergencies quickly and dispatch the right autonomous vehicle to execute response tasks.

A hybrid prediction framework merges grid and graph deep learning models to generate precise agent trajectories for autonomous navigation.

A system dispatches mobile energy delivery vehicles to re-energize electric vehicles at their current location.

A drone control module transfers operational authority to a local user device via established communication channels.

Sensor failure handling system switches autonomous vehicles from world to local coordination using camera data for lane tracking.

Autonomous vehicle control system predicts maximum acceleration, braking, and curvature capabilities to resolve safety versus merging efficiency trade-offs.