Semantic Obstacle Classification for Socially Aware Robot Navigation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Autonomous robots face challenges in navigating environments safely and socially, as existing path planning technologies often fail to account for dynamic obstacles and social considerations, leading to potential collisions and intrusive behavior.

Innovation Solution

A system that uses a planning engine to identify semantic classes of obstacles, applying specific planning policies to optimize route planning, avoiding collisions and respecting social boundaries by determining appropriate buffer zones and travel constraints for different types of objects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional path planning algorithms are used without semantic classification, then the system complexity is low, but the robot cannot differentiate between different types of obstacles and thus cannot apply appropriate planning policies, leading to potential collisions or socially intrusive behavior

Engineering Contradiction:
Improvecollision avoidanceVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments obstacles into different semantic classes (e.g., persons, pets, furniture, fragile objects) and assigns specific planning policies to each class. This segmentation allows the robot to differentiate between obstacle types and apply appropriate buffer zones and travel constraints, improving collision avoidance while maintaining manageable system complexity through modular policy design.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the parameter of obstacle representation from simple geometric detection to semantic classification with associated policy parameters. Each semantic class has defined parameters such as buffer zone distance, maximum approach speed, and travel cost weights, allowing the planning system to adapt its behavior based on the semantic type of each obstacle rather than treating all obstacles uniformly.

Inventive Principle:
Principle #35Parameter changes

2Ease of operation

If the robot applies uniform safety buffers to all obstacles, then the system is simple to operate, but it cannot respect social considerations for different obstacle types, leading to intrusive behavior toward persons and pets

Engineering Contradiction:
Improveoperation simplicityVSAvoidsocial consideration adaptability
Core Design Contradiction:
Ease of operationVSAdaptability or versatility

Solution Approach 1:

The patent applies local quality by assigning different planning policies to different semantic classes of obstacles. Each obstacle type receives localized treatment with appropriate buffer zones and constraints - for example, persons and pets receive larger buffer zones and higher travel costs, while furniture receives standard collision avoidance parameters. This allows the system to maintain simplicity in policy application while achieving adaptability to social considerations through semantic differentiation.

Inventive Principle:
Principle #3Local quality

3Measurement precision

If the system stores detailed information about each individual obstacle, then the navigation accuracy is high, but the data storage requirement and processing latency increase significantly

Engineering Contradiction:
Improvenavigation accuracyVSAvoiddata storage requirement
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent uses universal semantic classes that can represent multiple instances of similar obstacle types. Instead of storing detailed information about each individual obstacle, the system classifies obstacles into semantic categories (person, pet, furniture, fragile object) and applies pre-defined planning policies to each class. This universal approach maintains navigation accuracy by capturing the essential characteristics needed for safe and socially appropriate navigation while significantly reducing data storage requirements compared to storing detailed information about every individual obstacle.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP4043990B1Semantic obstacle recognition for path planning
Publication Date: 2024.03.27 GOOGLE LLC
  • EP4043990B1 patent drawingFigure 1A
  • EP4043990B1 patent drawingFigure 1B
  • EP4043990B1 patent drawingFigure 2

AI summary

Systems, methods, devices, and techniques for planning travel of an autonomous robot. A system identifies one or more obstacles that are located in proximity of at least a portion of a planned route for the autonomous robot. For each obstacle, the system: (i) determines a semantic class of the obstacle, including selecting the semantic class from a library that defines a set of multiple possible semantic classes for obstacles, and (ii) selects a planning policy for the obstacle that corresponds to the semantic class of the obstacle. The system can generate a trajectory along the at least the portion of the planned route using the selected planning policies. The robot can then initiate travel according to the trajectory.