Vehicle Guidance Algorithm for Automated Parking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for operating motor vehicles in navigation environments, such as parking environments, face challenges due to the high cost and limited sensor range and quality of sensor devices, which hinder the implementation of effective automated driving functions.
Innovation Solution
The use of a vehicle guidance algorithm trained with historical sensor data through machine learning, allowing for automated vehicle movement in navigation environments with a simpler sensor setup, and incorporating a collision avoidance algorithm for enhanced safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a complex sensor system is used for complete analytical evaluation of the vehicle's surroundings, then the reliability of automated driving function is improved, but the device complexity and cost increase significantly
Solution Approach 1:
The patent replaces the traditional mechanical sensor-based analytical evaluation system with a machine learning-based virtual sensor system. The vehicle guidance algorithm processes raw sensor data (from simpler sensors) to generate virtual sensor data that represents the evaluated situation, eliminating the need for complex sensor hardware while achieving reliable automated driving guidance.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between the simple sensor system and the automated driving control. This intermediary processes raw sensor inputs and generates meaningful situational understanding, enabling reliable automated guidance without requiring complex sensors.
2Ease of manufacture
If a simple sensor device is used for cost-effectiveness, then the device complexity is reduced, but the measurement precision and detection range become insufficient
Solution Approach 1:
The patent substitutes the physical sensor complexity with computational complexity. Instead of using high-precision sensors, it employs a machine learning-based vehicle guidance algorithm that enhances the measurement precision through data processing and pattern recognition, achieving accurate environmental understanding with simpler sensors.
Solution Approach 2:
The patent changes the parameters of the sensing system by transforming raw sensor data into virtual sensor data through machine learning processing. This parameter transformation enhances the effective precision and detection capabilities without changing the physical sensors.
3Device complexity
If machine learning is applied to process sensor data, then the device complexity is reduced, but the processing time and computational requirements increase
Solution Approach 1:
The patent applies preliminary action by pre-training the vehicle guidance algorithm with extensive training data before deployment. This pre-processing of the machine learning model enables rapid real-time inference during actual driving, reducing the processing time required during operation while maintaining simplicity of sensor setup.
Data Source
Figure 1
Figure 2
AI summary
The invention relates to a method for operating a motor vehicle (1) in a navigation surrounding area, in particular in a parking surrounding area. A digital navigation map, which describes a roadway grid, of the navigation surrounding area is used, and a navigation route to a destination contained in the navigation map is ascertained by a navigation system (2) of the motor vehicle (1) while taking into consideration the navigation map. Sensor data which describes the environment of the motor vehicle (1) is ascertained by means of a motor vehicle sensor device (6), and the navigation system (2) carries out a vehicle guidance algorithm which uses the sensor data as input data and is trained by means of a machine learning process using training data. The motor vehicle (1) is guided at least temporarily along the navigation route automatically according to at least one lateral and/or longitudinal guidance action ascertained by the vehicle guidance algorithm on the basis of the input data.