Monocular Sensor Fusion for Low-Power Autonomous Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies for providing fast, accurate, and reliable positional awareness to autonomous robots and mobile platforms are limited by the high cost, power consumption, and accuracy issues of existing methods such as RFID/WiFi, depth sensors, and visual approaches, which fail to meet the requirements for widespread adoption.
Innovation Solution
A monocular-auxiliary sensor system that utilizes a single operational camera with inertial measurement units and wheel odometry data to estimate movement and depth, offloading computational tasks to low-power sensors and processing units, enabling efficient image processing and energy conservation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If depth sensors are used for positional awareness, then measurement precision is improved, but use of energy increases and device cost increases
Solution Approach 1:
The patent combines monocular visual data with auxiliary sensor data (inertial sensors, wheel encoders, RFID readers) to create a fused positional estimate. This merging allows the system to achieve depth sensor-level accuracy without using power-hungry depth sensors, as the fusion of multiple low-power sensor streams compensates for the limitations of monocular vision.
Solution Approach 2:
The system uses a single camera that serves multiple functions: visual odometry, feature matching for localization, and scale estimation when combined with auxiliary sensors. This multi-functionality replaces dedicated depth sensors, reducing overall system power consumption while maintaining positional awareness accuracy.
2Ease of manufacture
If visual approaches are used for positional awareness, then device cost is reduced, but speed decreases and reliability worsens
Solution Approach 1:
The patent divides the computational workload by separating visual processing from auxiliary sensor processing. The monocular camera captures images at lower resolution and frame rates, while auxiliary sensors (inertial sensors, wheel encoders) operate independently at high frequencies. This segmentation allows each sensor type to operate in its optimal performance range, achieving both cost-effectiveness and high-speed responsiveness.
Solution Approach 2:
The patent introduces an intermediary fusion algorithm that combines visual odometry results with auxiliary sensor measurements. This intermediary processing layer reconciles the slower visual data with faster auxiliary sensor data, producing a unified high-speed positional estimate that maintains reliability while keeping device costs low.
3Ease of manufacture
If monocular cameras are used instead of depth sensors, then device cost is reduced, but measurement precision of depth decreases
Solution Approach 1:
The patent replaces optical depth sensing mechanisms (depth sensors, stereo cameras) with a computational approach using monocular visual odometry combined with mechanical auxiliary sensors (wheel encoders, inertial measurement units). This substitution uses the known kinematic model of the mobile platform and visual feature tracking to infer depth and scale, achieving accurate depth estimation without expensive depth-sensing hardware.
Solution Approach 2:
The system changes the parameters used for depth estimation by relying on temporal sequences of monocular images combined with auxiliary sensor measurements rather than direct optical depth sensing. By integrating visual flow, inertial measurements, and wheel encoder data over time, the system recovers scale and depth information that would otherwise require dedicated depth sensors, maintaining accuracy while reducing cost.
Data Source
AI summary
The described positional awareness techniques employing sensory data gathering and analysis hardware with reference to specific example implementations implement improvements in the use of sensors, techniques and hardware design that can enable specific embodiments to provide positional awareness to machines with improved speed and accuracy. The sensory data are gathered from an operational camera and one or more auxiliary sensors.