Research Needs for Automobile Safety
Electronics represents a significant portion of a vehicle's cost. Electronics control everything from the operation of the engine and brakes, to the control of passenger safety, fuel efficiency, vehicle stability comfort, and security. To accomplish these tasks, vehicles are equipped with a vast array of sensors, actuators, cameras, satellite navigation devices, communications equipment, and advanced electronic components. Electronic control units (ECUs) are used to manage all these computing resources, leading to complex data processing systems that must reliably communicate with and integrate multiple elements under harsh conditions over the life of the vehicle in order to provide a safe and efficient driving experience.
As manufacturers compete to provide the most advanced features and capabilities to their vehicles at an affordable cost, new technologies, with less history of prior use, are being adopted at a rapid pace. When these features and capabilities are implemented improperly, safety issues and recalls will occur [5]-[7], [13]-[17]. More rapid and effective qualification methods are needed to reveal potential risks associated with new sensors, active and passive components, engine control units, and other systems with hardware/software interactions [1] [10]. Research is also needed to improve lot acceptance and screening techniques to identify faulty electronic components before they are assembled into vehicles. For instance, failure mechanisms related to wire bonding, high density printed circuit board technologies, and advanced capacitors and energy storage devices must be researched.
One major upcoming change in automotive electronics, which will occur in January 2016, is the elimination of lead (Pb) in solder used to connect electronic components to printed wiring boards under the End of Life Vehicle (ELV) Directive 2011/37/EU in the European Union. The globalization of the automotive supply chain ensures that this requirement will impact automotive electronics in all countries. Tin-lead solder has been the standard for assembly of electronics for more than 60 years, and test methods as well as reliability and performance expectations are built on this historical knowledge. This is pushing automobile manufacturers to evaluate and select a replacement lead-free solder. While consumer-based electronics have converted to tin-silver-copper solder, performance and reliability concerns, particularly in harsh application environments, have held back the automotive industry. The required engineering material change will impact manufacturing, testing, and automotive reliability and safety. The updated ELV directive also calls for the elimination of lead on component terminal finishes. The elimination of lead in consumer applications has resulted in the wide adoption of tin and tin-based alloys that are susceptible to tin whisker formation. The growth of tin whiskers has been known to cause electrical shorts and electronic failures [8] [9] [22]-[25].
Battery safety and reliability are critical for the development of hybrids and electric vehicles [3]. In addition, the evaluation of battery life through state of charge and state of health estimation is essential for the management of batteries. State of health can be estimated using ultrasonic signal data and changes of other physical attributes as inputs with on-board sensors. Research is also needed in the area of Li-ion batteries, where failures have resulted in vehicle fires.
A key problem with the electronics used in automotive applications is intermittent failures that can result in no-fault-founds when taken to the dealer [18]-[21]. These events add unnecessary cost and create dissatisfaction over the automotive distribution supply chain. Storage of event details with sufficient granularity will help to resolve these situations. Research is also needed to detect anomalies and trends in multivariate data streams. These capabilities could be beneficial for fault diagnostics of vehicles [2] during qualification as well as operation. However, the selection of parameters and criteria for anomaly detection as well as sensor optimization requires additional research.
Electronic control systems in automobiles should be equipped with the same kinds of fault detection and prognostics capabilities that have already been implemented in mechanical systems, in order to ensure that electronic malfunctions do not lead to safety lapses. Research is needed to decompose complex electronic systems into multiple independent subsystems which can be monitored in-situ. Fault propagation among these subsystems should be studied to evaluate the impact of subsystems' faults on the performance of the whole system. Research should be conducted in fusion technologies which leverage the advantages of data-driven [11] and physics-of-failure [12] prognostic methods to handle diagnostic and prognostic uncertainties that arise from component unit-to-unit variation, measurement noise, dynamical loading conditions, and system modeling inaccuracy. Forecasted maintainability can then be accomplished to reduce vehicle failure rates and increase operational safety.
Supply chain management issues have been at the heart of many past automotive failures. Research is needed to study reliability capability assessment methods as part of the supplier selection and part selection processes. Counterfeit electronic components are a growing concern related to the electronic parts supply chain [4], and methods for prevention must be researched. Finally, research is needed to improve the design, development, and processing of flex and rigid/flex substrates and connectors, as well as design of experiments for their integrity during a crash and how they can fail under impact.
References