Search for: All records

Machine-learning (ML) technology has been a key enabler in the push towards realizing ever more sophisticated autonomous-driving features. In deploying such technology, the automotive industry has relied heavily on using "black-box" software and hardware components that were originally intended for non-safety-critical contexts, without a full understanding of their real-time capabilities. A prime example of such a component is CUDA, which is fundamental to the acceleration of ML algorithms using NVIDIA GPUs. In this paper, evidence is presented demonstrating that CUDA can cause unbounded task delays. Such delays are the result of CUDA’s usage of synchronization mechanisms in the POSIX thread (pthread) library, so the latter is implicated as a delay-prone component as well. Such synchronization delays are shown to be the source of a system failure that occurred in an actual autonomous vehicle system during testing at WeRide. Motivated by these findings, a broader experimental study is presented that demonstrates several real-time deficiencies in CUDA, the glibc pthread library, Linux, and the POSIX interface of the safety-certified QNX Operating System for Safety. Partial mitigations for these deficiencies are presented and further actions are proposed for real-time researchers and developers to integrate more complete mitigations. 

Prussian blue analogs (PBAs) are an important material class for aqueous electrochemical separations and energy storage owing to their ability to reversibly intercalate monovalent cations. However, incorporating interstitial [Formula: see text] molecules in the ab initio study of PBAs is technically challenging, though essential to understanding the interactions between interstitial water, interstitial cations, and the framework lattice that affect intercalation potential and cation intercalation selectivity. Accordingly, we introduce and use a method that combines the efficiency of machine-learning models with the accuracy of ab initio calculations to elucidate mechanisms of (1) lattice expansion upon intercalation of cations of different sizes, (2) selectivity bias toward intercalating hydrophobic cations of large size, and (3) semiconductor–conductor transitions from anhydrous to hydrated lattices. We analyze the PBA nickel hexacyanoferrate [[Formula: see text]] due to its structural stability and electrochemical activity in aqueous electrolytes. Here, grand potential analysis is used to determine the equilibrium degree of hydration for a given intercalated cation (Na[Formula: see text], K[Formula: see text], or Cs[Formula: see text]) and [Formula: see text] oxidation state based on pressure-equilibrated structures determined with the aid of machine learning and simulated annealing. The results imply new directions for the rational design of future cation-intercalation electrode materials that optimize performance in various electrochemical applications, and they demonstrate the importance of choosing an appropriate calculation framework to predict the properties of PBA lattices accurately.