Smart Buffer Memory Device for Automotive Sensor Data Prioritization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional memory devices are unable to provide the required data processing and transmission speeds for the large volumes of sensing data from automotive sensors like image sensors, radars, and lidars without increasing costs or physical size, and they lack sufficient data transmission channels.
Innovation Solution
A memory device with a smart buffer that classifies sensing data by importance using a priority setting unit and allocates channels accordingly, utilizing a neural network to determine data weight and dynamically assign channels to high-speed or low-speed memory areas based on priority, allowing for efficient data processing and transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional memory devices are used to handle large volumes of sensing data, then data storage capacity is sufficient, but data processing and transmission speeds are insufficient
Solution Approach 1:
The memory area is divided into first and second memory areas, and channels are segmented into first and second channel groups. High-priority sensing data is routed through the first channel group to the first memory area for fast processing, while low-priority data uses the second channel group and second memory area. This segmentation enables differential data processing speeds within the same memory device.
Solution Approach 2:
The channel controller dynamically allocates channels to channel groups based on the priority of sensing data. The priority setting unit determines data priority using weights from neural network operations, and the channel controller adjusts channel assignment in real-time. This dynamic allocation optimizes data transmission speed according to actual data importance without requiring fixed hardware configurations.
2Productivity
If more data transmission channels are provided to increase data processing speed, then data transmission capacity improves, but device complexity and cost increase
Solution Approach 1:
The channel controller serves multiple functions: it manages channel allocation, determines data priority through the priority setting unit, and routes data to appropriate channel groups. The same channel controller handles both high-priority and low-priority data, adjusting its behavior based on data characteristics rather than requiring separate controllers for each function. This multi-functionality reduces overall device complexity.
Solution Approach 2:
The priority setting unit automatically determines data priority using weights obtained from neural network operations performed on the sensing data itself. The system self-regulates channel allocation based on data characteristics without external intervention, reducing the need for complex external control mechanisms and simplifying the overall system architecture.
3Reliability
If data is transmitted without priority classification, then transmission channel usage is simple, but important sensing data may not receive sufficient transmission bandwidth
Solution Approach 1:
The priority setting unit continuously monitors sensing data and adjusts priority classification based on weights from neural network operations. The channel controller receives feedback about data priority and dynamically adjusts channel allocation accordingly. This feedback mechanism ensures that important sensing data consistently receives adequate transmission bandwidth while maintaining a relatively simple allocation structure.
Data Source
AI summary
A memory device comprises a smart buffer, and a memory area divided into a first memory area and a second memory area, wherein the smart buffer comprises a priority setting unit configured to receive a sensing data and a corresponding weight from a controller, determine a priority of the sensing data based on the weight, and classify the sensing data as one of first priority sensing data and second priority sensing data, and a channel controller configured to allocate at least one channel selected from among a plurality of channels to a first channel group, allocate at least another channel selected from among the plurality of channels to a second channel group, assign the first channel group to process the first priority sensing data in relation to the first memory area, and assign the second channel group to process the second priority sensing data in relation to the second memory area, wherein a number of data input/output (I/O) pins connected to the first channel group is greater than a number of data I/O pins connected to the second channel group, wherein the memory area includes at least one memory chip, wherein the at least one memory chip includes a first chip having a first metal pad and a cell region and a second chip having a second metal pad and a peripheral circuit region, and the first chip and the second chip are vertically connected to each other by the first metal pad and the second metal pad.