LiDAR Point Cloud Storage Format Using Dynamic Bit Widths
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for storing point cloud data from LiDAR sensors in autonomous vehicles are inefficient, as they either require expensive hardware implementations or fail to consider sensor properties, leading to unnecessary storage space usage.
Innovation Solution
A user-defined storage format using flexible bitwise vectors is employed, where each vector includes multiple data segments and a metadata segment, with bit widths dynamically adjusted based on LiDAR sensor hardware properties, optimizing storage space by inversely proportional bit width adjustment according to the number of laser lines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard point cloud data formats are used for storage, then data precision is maintained, but storage space is wasted
Solution Approach 1:
The patent applies local quality by using different bit widths for different fields within the point cloud data structure. Specifically, the x, y, z coordinate fields use 16 bits each, while the intensity field uses only 8 bits, and the time field uses 32 bits. This localized optimization of data precision per field reduces overall storage requirements while maintaining necessary precision for each specific attribute.
Solution Approach 2:
The patent changes the parameter of bit width from a fixed standard value to a optimized value based on actual sensor capabilities. By adjusting the bit width parameters of different data fields according to the LiDAR sensor's resolution and measurement range, the system achieves compact storage without sacrificing the precision actually provided by the sensor hardware.
2Quantity of substance
If lossless or lossy algorithms are used to reduce point cloud data size, then storage space is reduced, but hardware implementation cost increases
Solution Approach 1:
The patent changes the data representation parameters (bit widths) rather than applying complex compression algorithms. This parameter-based optimization achieves storage reduction through simple data structure design rather than computationally intensive compression techniques, thereby reducing hardware implementation complexity and cost.
Solution Approach 2:
The patent uses a simple, lightweight data structure format that can be easily implemented and processed without requiring expensive hardware compression modules. The optimized binary format is straightforward to implement in standard processors, avoiding the need for specialized hardware acceleration.
3Ease of operation
If fixed bit width is used for all data fields, then data structure simplicity is maintained, but storage efficiency decreases
Solution Approach 1:
The patent implements local quality by assigning different bit widths to different fields based on their specific precision requirements. The data structure remains simple and regular with fixed positions for each field, but the bit width parameter varies locally for each field (x: 16 bits, y: 16 bits, z: 16 bits, intensity: 8 bits, time: 32 bits), achieving both simplicity and efficiency.
Data Source
AI summary
In one embodiment, an exemplary computer-implemented method of storing point cloud data in an autonomous driving vehicle can include the operations of receiving raw point cloud data from a LiDAR sensor mounted on the autonomous driving vehicle, the raw point cloud data representing cloud data points acquired in response to laser beams emitted at a given angle; retrieving configuration information of the LiDAR sensor, the configuration information including at least a number of laser lines of the LiDAR sensor. The method further includes the operations of constructing, based on the configuration information, a data structure that includes a data entry for each of the cloud data points, the data entry including multiple fields for storing attributes of the cloud data point, each field having a bit width determined based on the configuration information using a predetermined algorithm; and writing the cloud data points to a storage medium using the data structure.