Time-Aware Occupancy Grid Mapping for Dynamic Robot Environments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional occupancy mapping technologies fail to consider the dynamic changes in environments, leading to inaccurate occupancy probability assessments due to the equal weighting of recent and outdated sensor observations, which is particularly problematic in dynamic scenarios.
Innovation Solution
Implementing a time-aware occupancy mapping system that uses Regression to Unknown (RTU) methodology, where each cell's occupancy probability is stored with a timestamp, and the probability is updated based on both the stored value and timestamp, with different regression factors for static and movable objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional occupancy mapping technologies use equal weighting for all sensor observations, then the system is simple to implement, but the occupancy probability assessment becomes inaccurate in dynamic environments
Solution Approach 1:
The system dynamically adjusts the weighting of occupancy probabilities based on temporal information. Instead of using static equal weighting, the patent applies time-dependent regression factors that adapt to environmental changes, making the mapping system responsive to dynamic conditions while maintaining reasonable complexity through structured temporal modeling
Solution Approach 2:
The patent changes the parameter of occupancy probability weighting from a static equal value to a dynamic value that depends on time elapsed since last observation. By introducing temporal parameters and regression factors that vary with time, the system improves accuracy in dynamic environments without requiring complete system redesign
2Speed
If the system stores only recent sensor observations with high weight, then the occupancy map responds quickly to changes, but outdated but still relevant information is lost
Solution Approach 1:
The system implements periodic updating of occupancy probabilities with time-aware regression. Instead of discarding old data or giving it equal weight, the patent applies systematic periodic regression at each time step, allowing information to be retained and appropriately weighted based on its age, thus preventing information loss while maintaining responsive updates
Solution Approach 2:
The patent incorporates feedback mechanisms where the system continuously monitors the age of observations and adjusts weighting accordingly. The temporal regression process provides feedback loops that prevent both premature discarding of useful historical data and excessive reliance on outdated information, balancing responsiveness with information retention
3Measurement precision
If the system applies different regression factors for static and movable objects, then the occupancy accuracy for dynamic objects improves, but the computational complexity increases
Solution Approach 1:
The patent segments the environment into static and dynamic regions, applying different regression factors to each segment. This segmentation allows the system to handle complex dynamic objects with appropriate temporal weighting while maintaining simpler processing for static areas, thereby managing computational complexity through spatial partitioning
Solution Approach 2:
The system applies local quality by using different regression factors in different spatial locations based on whether objects are static or movable. Instead of uniformly complex processing everywhere, the patent tailors the temporal regression behavior to local characteristics, improving accuracy where needed while reducing unnecessary computational overhead in static regions
Data Source
AI summary
A time-aware occupancy mapping using regression to unknown (“RTU”) analysis, and an apparatus to dynamically allocate occupancy probability to a cell in an environment to thereby form a time-aware occupancy map of the environment are disclosed. The apparatus includes a memory circuitry in communication with a processor circuitry, the memory circuitry configured to receive and store probability information from the processor circuitry and to store the probability value and its corresponding time stamp at a probability table. The processor circuitry may be configured to, among others: (1) receive occupancy information, the occupancy information defining whether a first of a plurality of cells (102, 104, 106 and 108) in the environment is occupied; (2) determine a first probability value that the first cell is occupied at a first point in time; (3) direct the first probability value and its corresponding timestamp to the memory circuitry to store; (4) determine a second probability value that the first cell is occupied at a second point in time, the second point in time defined by a lapsed interval from the first point in time to the second point in time; and (5) update the memory circuitry to store the second probability value and its corresponding timestamp.