Hierarchical Map Representation for Task-Specific Autonomous Operations

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing autonomous agent systems face challenges in efficiently storing and retrieving task-specific representations of complex environments, leading to compute, storage, and latency overheads, which hinder real-time performance and power efficiency.

Innovation Solution

A map representation system that provides a unified and scalable space representation methodology with hierarchical access to representations ranging from highly granular surfaces to fine granular point clouds, enabling intelligent data storage and retrieval based on semantic requirements and context-specific tasks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If comprehensive environmental data is stored for all tasks, then representation accuracy is improved, but storage requirements and compute overhead increase

Engineering Contradiction:
Improverepresentation accuracyVSAvoidstorage requirements
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent segments environmental representations into multiple granularities (coarse-grained voxel maps for navigation, fine-grained point clouds for manipulation). Each granularity level stores only the data necessary for specific task types, avoiding the need to store all possible detail levels for all tasks. This segmentation resolves the contradiction by providing high accuracy only where needed while reducing overall storage requirements.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system applies local quality by storing high-resolution data only in regions and at granularities relevant to current tasks. For example, fine-grained point cloud data is stored only for objects requiring manipulation, while navigation areas use coarser voxel representations. This localized application of high quality data reduces total storage while maintaining accuracy for task-critical regions.

Inventive Principle:
Principle #3Local quality

2Reliability

If high-resolution representations are stored for all tasks, then task performance is improved, but processing time and latency increase

Engineering Contradiction:
Improvetask performanceVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent implements dynamic representation retrieval where the system selects appropriate granularity levels and data types based on current task requirements. For navigation tasks, coarse-grained voxel maps are retrieved for fast processing, while manipulation tasks trigger retrieval of fine-grained point cloud data. This dynamic adaptation resolves the contradiction by optimizing processing time for each task type while maintaining high performance where needed.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes representation parameters (granularity level, data type, resolution) based on task context. The same environmental space can be represented at different parameter levels depending on the operation required. This parameter adaptation allows the system to achieve high task performance with minimal processing time by using only the necessary level of detail for each specific operation.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If multiple representation types are maintained for different tasks, then task-specific accuracy is improved, but system complexity increases

Engineering Contradiction:
Improvetask-specific accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent creates a universal hierarchical representation system where a single integrated data structure can serve multiple task types through selective access to different granularity levels. The same underlying representation framework supports both navigation and manipulation tasks by providing appropriate data resolutions, eliminating the need for completely separate representation systems for each task type. This universality reduces system complexity while maintaining task-specific accuracy.

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

Solution Approach 2:

The system implements nested representations where fine-grained data structures are embedded within coarser-grained frameworks. Point cloud data for detailed manipulation is nested within the broader voxel map structure for navigation. This nesting allows the system to maintain multiple representation types in a unified, organized manner, reducing overall system complexity while preserving task-specific accuracy through selective access to appropriate nested levels.

Inventive Principle:
Principle #7Nested doll (Nesting)

4Adaptability or versatility

If detailed environmental maps are stored, then autonomous agent capability is improved, but power consumption increases

Engineering Contradiction:
Improveautonomous agent capabilityVSAvoidpower consumption
Core Design Contradiction:
Adaptability or versatilityVSUse of energy by moving object

Solution Approach 1:

The patent applies partial action by storing and processing only the necessary level of detail for current autonomous operations. Instead of continuously maintaining and processing all possible representation details, the system retrieves and processes only the partial data set required for the current task (e.g., coarse navigation data during movement, fine-grained data only when interaction is needed). This partial processing significantly reduces power consumption while maintaining full autonomous capability when needed.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP4015997A1Map representation system and computer-readable medium
Publication Date: 2022.06.22 INTEL CORP
  • EP4015997A1 patent drawingFigure 1
  • EP4015997A1 patent drawingFigure 2
  • EP4015997A1 patent drawingFigure 3

AI summary

Various aspects are related to a map representation system including one or more processors configured to: obtain operation task data, the operation task data indicating one or more operation tasks associated with an operation of one or more autonomous agents within a corresponding operating space; and obtain operation task specific map data as a function of the operation task data, the operation task specific map data representing the operating space corresponding to the one or more operation tasks.