Model-Based Firmware Generation for Acquisition Systems Using Heterogeneous Hardware

Abstract

High-performance sensing and control systems have an important role in Industry 4.0. However, with the current solutions, the development effort is high and requires specialized skills in electronic engineering. Therefore, a model-based approach on control and signal processing systems using affordable heterogeneous hardware is proposed. In this work, a model-based code generator is developed to abstract the user from the actual software implementation. Starting from a combined model of a timing diagram and an embedded platform, a model transformation is used to automatically generate functional acquisition firmware. This firmware generator enables system engineers without deep software and hardware knowledge to set up complex control systems. Furthermore, it equips software engineers with a solid framework for faster development.

Automated Firmware Generation for Compressive Sensing on Heterogeneous Hardware

“…This paper demonstrates that the proposed framework from our previous work [ 1 ] is generic enough to support complex sampling strategies, such as CS, and as such unleashes the power of heterogeneous platforms to the larger public. In Figure 1 , an overview of the current status of the framework is presented.…”

“…As a first test, the model-based code generator is used to generate a traditional Shannon-conform acquisition sequence for the ADS8556. This firmware has been thoroughly evaluated in previous work [ 1 ] and will serve as the ground truth for the CS firmware to be compared with. The development of the ground truth firmware using the code generator took about three hours, including testing.…”

“…This is parsed into the eclipse modeling framework. From here, the acquisition sequence is automatically processed to generate a functional firmware in agreement with the specified acquisition sequence and heterogeneous platform [ 1 ].…”

“…With the implementation of minimalistic platform-specific libraries, the framework maintains a good level of portability, platform independency and maintainability. Previous research has demonstrated that a generic cross-platform development of acquisition systems respecting the modeled timing requirements can be achieved [ 1 ]. This paper explores whether the framework is generic enough to support a complex sampling scheme, i.e., compressive sensing (CS).…”