Search for: All records

The paper discussesalgorithmic priority inversionin mission-critical machine inference pipelines used in modern neural-network-based perception subsystems and describes a solution to mitigate its effect. In general,priority inversionoccurs in computing systems when computations that are less important are performed together with or ahead of those that are more important. Significant priority inversion occurs in existing machine inference pipelines when they do not differentiate between critical and less critical data. We describe a framework to resolve this problem and demonstrate that it improves a perception system's ability to react to critical inputs, while at the same time reducing platform cost. 

Abstract Advances in deep learning have revolutionized cyber‐physical applications, including the development of autonomous vehicles. However, real‐world collisions involving autonomous control of vehicles have raised significant safety concerns regarding the use of deep neural networks (DNNs) in safety‐critical tasks, particularly perception. The inherent unverifiability of DNNs poses a key challenge in ensuring their safe and reliable operation. In this work, we propose perception simplex ( ), a fault‐tolerant application architecture designed for obstacle detection and collision avoidance. We analyse an existing LiDAR‐based classical obstacle detection algorithm to establish strict bounds on its capabilities and limitations. Such analysis and verification have not been possible for deep learning‐based perception systems yet. By employing verifiable obstacle detection algorithms, identifies obstacle existence detection faults in the output of unverifiable DNN‐based object detectors. When faults with potential collision risks are detected, appropriate corrective actions are initiated. Through extensive analysis and software‐in‐the‐loop simulations, we demonstrate that provides deterministic fault tolerance against obstacle existence detection faults, establishing a robust safety guarantee. 

Perception of obstacles remains a critical safety concern for autonomous vehicles. Real-world collisions have shown that the autonomy faults leading to fatal collisions originate from obstacle existence detection. Open source autonomous driving implementations show a perception pipeline with complex interdependent Deep Neural Networks. These networks are not fully verifiable, making them unsuitable for safety-critical tasks. In this work, we present a safety verification of an existing LiDAR based classical obstacle detection algorithm. We establish strict bounds on the capabilities of this obstacle detection algorithm. Given safety standards, such bounds allow for determining LiDAR sensor properties that would reliably satisfy the standards. Such analysis has as yet been unattainable for neural network based perception systems. We provide a rigorous analysis of the obstacle detection s 

Abstract Consensus on the cause of recent midlatitude circulation changes toward a wavier manner in the Northern Hemisphere has not been reached, albeit a number of studies collectively suggest that this phenomenon is driven by global warming and associated Arctic amplification. Here, through a fingerprint analysis of various global simulations and a tropical heating-imposed experiment, we suggest that the suppression of tropical convection along the Inter Tropical Convergence Zone induced by sea surface temperature (SST) cooling trends over the tropical Eastern Pacific contributed to the increased summertime midlatitude waviness in the past 40 years through the generation of a Rossby-wave-train propagating within the jet waveguide and the reduced north-south temperature gradient. This perspective indicates less of an influence from the Arctic amplification on the observed mid-latitude wave amplification than what was previously estimated. This study also emphasizes the need to better predict the tropical Pacific SST variability in order to project the summer jet waviness and consequent weather extremes. 

We present the first Greenlandic tree ring oxygen isotope record (δ18OGTR), derived from four birch trees collected from the Qinguadalen Valley in southwestern Greenland in 1999. Our δ18O record spans from 1950–1999 and is significantly and positively correlated with winter ice core δ18O from southern Greenland. δ18OGTR records are positively correlated with southwestern Greenland January–August mean temperatures. North Atlantic Oscillation (NAO) reconstructions have been developed from a variety of proxies, but never with Greenlandic tree rings, and our δ18OGTR record is significantly correlated with NAO (r = −0.64), and spatial correlations with sea-level pressure indicate a classic NAO pressure seesaw pattern. These results may facilitate a longer NAO reconstruction based on long time series of tree ring δ18O records from Greenland, provided that subfossil wood can be found in areas vacated by melting glaciers. 

Availability of water for irrigated crops is driven by climate and policy, as moderated by public priorities and opinions. We explore how climate and water policy interact to influence water availability for cannabis (Cannabis sativa), a newly regulated crop in California, as well as how public discourse frames these interactions. Grower access to surface water covaries with precipitation frequency and oscillates consistently in an energetic 11–17 year wet-dry cycle. Assessing contemporary cannabis water policies against historic streamflow data showed that legal surface water access was most reliable for cannabis growers with small water rights (<600 m3) and limited during relatively dry years. Climate variability either facilitates or limits water access in cycles of 10–15 years—rendering cultivators with larger water rights vulnerable to periods of drought. However, news media coverage excludes growers’ perspectives and rarely mentions climate and weather, while public debate over growers’ irrigation water use presumes illegal diversion. This complicates efforts to improve growers’ legal water access, which are further challenged by climate. To promote a socially, politically, and environmentally viable cannabis industry, water policy should better represent growers’ voices and explicitly address stakeholder controversies as it adapts to this new and legal agricultural water user. 

The paper discusses algorithmic priority inversion in mission-critical machine inference pipelines used in modern neural-network-based cyber-physical applications and develops a scheduling solution to mitigate its effect. In general, priority inversion occurs in real-time systems when computations that are of lower priority are performed together with or ahead of those that are of higher priority. 1 In current machine intelligence software, significant priority inversion occurs on the path from perception to decision-making, where the execution of underlying neural network algorithms does not differentiate between critical and less critical data. We describe a scheduling framework to resolve this problem and demonstrate that it improves the system's ability to react to critical inputs, while at the same time reducing platform cost.