Autonomous Navigation Image Processing System

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Autonomous moving platforms, such as robots and drones, face challenges in accurately navigating and mapping their environment using existing sensor systems, particularly in real-time obstacle avoidance and path planning, due to limitations in processing visual and range data effectively.

Innovation Solution

A system employing a CPU with image processing capabilities, utilizing a combination of RGBD devices and LiDAR sensors, and a Graphics Processing Unit (GPU) for iterative optimization and rendering, creates a rendered model of the environment, updating it based on sensor data to accurately determine the platform's position and orientation, and enhance navigation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If existing sensor systems are used for environmental mapping and navigation, then the platform can perform basic autonomous navigation, but the accuracy and real-time performance of obstacle avoidance and path planning are insufficient

Engineering Contradiction:
Improveenvironmental mapping accuracyVSAvoidreal-time processing speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent segments the monolithic sensor processing approach into separate specialized sensors (RGBD camera for visual features, LiDAR for depth information) and processes their data through distinct pipelines that are later integrated. This segmentation allows each sensor type to be optimized for its specific function while maintaining overall system real-time performance.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent merges data from multiple sensor types (RGBD imaging sensor and LiDAR range sensor) to create a unified environmental model. By combining visual features with depth information, the system achieves higher mapping accuracy while using complementary sensor strengths to maintain real-time processing capability.

Inventive Principle:
Principle #5Merging (Combining)

2Measurement precision

If multiple sensor types are combined for better environmental perception, then the mapping accuracy improves, but the system complexity increases

Engineering Contradiction:
Improveenvironmental perception accuracyVSAvoidsensor system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent implements a unified processing framework that handles multiple sensor types through common algorithms and data structures. The environmental model and processing pipeline are designed to accommodate different sensor inputs universally, reducing the complexity that would otherwise arise from separate processing chains for each sensor type.

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

Solution Approach 2:

The patent introduces an intermediary processing layer that standardizes data from different sensor types into a common format before integration. This intermediary representation layer simplifies the complexity of directly fusing heterogeneous sensor data by providing a standardized interface between diverse sensors and the core processing algorithms.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If advanced processing algorithms are used to improve navigation accuracy, then the path planning precision improves, but the computational time and processing load increase

Engineering Contradiction:
Improveposition and orientation determination accuracyVSAvoidcomputational processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary processing of sensor data to extract key features and pre-compute environmental models before full navigation decisions are required. By preparing data structures and environmental representations in advance, the system reduces the computational burden during critical real-time navigation and path planning moments.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements a hierarchical processing approach where essential navigation functions use computationally efficient algorithms for real-time response, while more accurate but computationally intensive algorithms are applied selectively for non-critical updates or when computational resources are available. This partial application of advanced algorithms maintains real-time performance while improving accuracy where feasible.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3311363B1System and method for processing captured images
Publication Date: 2020.08.05 ROSEMOUNT AEROSPACE INC
  • EP3311363B1 patent drawingFigure 1
  • EP3311363B1 patent drawingFigure 2

AI summary

An image processing system configured to process perceived images of an environment includes a central processing unit (CPU) including a memory storage device having stored thereon a computer model of the environment, at least one sensor configured and disposed to capture a perceived environment including at least one of visual images of the environment and range data to objects in the environment, and a rendering unit (RU) configured and disposed to render the computer model of the environment forming a rendered model of the environment. The image processing system compares the rendered model of the environment to the perceived environment to update the computer model of the environment.