Abstract: Many Embedded Systems are indeed Software Based Control Systems (SBCSs), that is control systems whose controller consists of control software running on a microcontroller device. This motivates investigation on Formal Model Based Design approaches for control software. Given the formal model of a plant as a Discrete Time Linear Hybrid System and the implementation specifications (that is, number of bits in the Analog-to-Digital (AD) conversion) correct-by-construction control software can be automatically generated from System Level Formal Specifications of the closed loop system (that is, safety and liveness requirements), by computing a suitable finite abstraction of the plant.
With respect to given implementation specifications, the automatically generated code implements a time optimal control strategy (in terms of set-up time), has a Worst Case Execution Time linear in the number of AD bits $b$, but unfortunately, its size grows exponentially with respect to $b$. In many embedded systems, there are severe restrictions on the computational resources (such as memory or computational power) available to microcontroller devices.
This paper addresses model based synthesis of control software by trading system level non-functional requirements (such us optimal set-up time, ripple) with software non-functional requirements (its footprint). Our experimental results show the effectiveness of our approach: for the inverted pendulum benchmark, by using a quantization schema with 12 bits, the size of the small controller is less than 6% of the size of the time optimal one.

Abstract: Many software as well digital hardware automatic synthesis methods define the set of implementations meeting the given system specifications with a boolean relation K. In such a context a fundamental step in the software (hardware) synthesis process is finding effective solutions to the functional equation defined by K. This entails finding a (set of) boolean function(s) F (typically represented using OBDDs, Ordered Binary Decision Diagrams) such that: 1) for all x for which K is satisfiable, K(x, F(x)) = 1 holds; 2) the implementation of F is efficient with respect to given implementation parameters such as code size or execution time. While this problem has been widely studied in digital hardware synthesis, little has been done in a software synthesis context. Unfortunately the approaches developed for hardware synthesis cannot be directly used in a software context. This motivates investigation of effective methods to solve the above problem when F has to be implemented with software. In this paper we present an algorithm that, from an OBDD representation for K, generates a C code implementation for F that has the same size as the OBDD for F and a WCET (Worst Case Execution Time) at most O(nr), being n = |x| the number of arguments of functions in F and r the number of functions in F.

Abstract: In this paper we show that statistical properties of the transition graph of a system to be verified can be exploited to improve memory or time performances of verification algorithms. We show experimentally that protocols exhibit transition locality. That is, with respect to levels of a breadth-first state space exploration, state transitions tend to be between states belonging to close levels of the transition graph. We support our claim by measuring transition locality for the set of protocols included in the Mur$\varphi$ verifier distribution. We present a cache-based verification algorithm that exploits transition locality to decrease memory usage and a disk-based verification algorithm that exploits transition locality to decrease disk read accesses, thus reducing the time overhead due to disk usage. Both algorithms have been implemented within the Mur$\varphi$ verifier. Our experimental results show that our cache-based algorithm can typically save more than 40% of memory with an average time penalty of about 50% when using (Mur$\varphi$) bit compression and 100% when using bit compression and hash compaction, whereas our disk-based verification algorithm is typically more than ten times faster than a previously proposed disk-based verification algorithm and, even when using 10% of the memory needed to complete verification, it is only between 40 and 530% (300% on average) slower than (RAM) Mur$\varphi$ with enough memory to complete the verification task at hand. Using just 300 MB of memory our disk-based Mur$\varphi$ was able to complete verification of a protocol with about $10^9$ reachable states. This would require more than 5 GB of memory using standard Mur$\varphi$.