Browse Topic: Telemetry
The reliability of Drive Unit (DU) oil pumps is critical to the performance and safety of electric vehicles, as these pumps provide essential lubrication and thermal management. In modern EV architectures, real-time health monitoring of these pumps typically relies on indirect signals than dedicated sensing hardware, a design choice optimized for cost, weight, and system complexity. This makes early fault detection a non-trivial challenge. To address this limitation, we present a novel, data-driven anomaly detection framework that leverages large-scale customer fleet telemetry and advanced machine learning to identify incipient pump degradation that traditional diagnostic methods often fail to capture. Specifically, we develop an XGBoost regression model trained on time-series features—including commanded pump speed, oil temperature, and historical pump current—to predict expected current behavior under nominal conditions. Deviations are quantified using the Mean Absolute Percentage Error (MAPE) between predicted and actual currents, providing a continuous and interpretable measure of anomaly severity. A fully automated pipeline ingests daily telemetry, performs session segmentation, executes predictive modeling, and records anomaly outcomes in backend databases for continuous monitoring and engineering review. The proposed framework enables continuous, fleet-wide predictive maintenance of DU oil pumps. It improves early detection of degradation, reduces vehicle downtime, enhances safety, and increases customer satisfaction. More broadly, it highlights the potential of large-scale data analytics and machine learning to advance predictive maintenance and reliability in electric vehicle (EV) systems.
Cars that are more connected, equipped with more sensors than ever before, should make proactive maintenance somewhat easy and reliable. Drivers could have lower repair costs and fewer breakdowns overall if the automotive industry shifts away from a periodic maintenance mindset towards data-driven proactive services. But the automakers themselves would also win with a massive drop in recalls. A connected car requires accurate telemetry sensors to track details such as temperature (in the engine, battery, and cabin), pressure (in tires, fuel, and oil), electrical current/voltage, and vibration patterns to detect problems before they become failures. The vehicle also needs to be able to combine telemetry, diagnostics, metadata, and service records in one platform and then make sense of it all. “Poor data integration kills even the best analytics,” according to Upstream co-founder and CTO Yonatan Appel.
Virtualization features such as digital twins and virtual patching can accelerate development and make commercial vehicles more agile and secure. There is one sure-fire way to secure commercial vehicles from cyber-attacks. “You just remove the connectivity,” quipped Brandon Barry, CEO of Block Harbor Cybersecurity and the moderator of a panel session on “cybersecurity of virtual machines” at the SAE COMVEC 2024 conference in Schaumburg, Illinois. Obviously, that train has left the station - commercial vehicles of all types, including trains, are only becoming more automated and connected, which increases the risks for cyber-attacks. “We have very connected vehicles, so attacks can be posed not just through powertrain solutions but also through telemetry, infotainment systems connected to different applications and services, and also through cloud platforms,” said Trisha Chatterjee, current product support and data specialist for fuel cell and hydrogen technology at Accelera by Cummins.
Hypersonic platforms provide a challenge for flight test campaigns due to the application's flight profiles and environments. The hypersonic environment is generally classified as any speed above Mach 5, although there are finer distinctions, such as “high hypersonic” (between Mach 10 to 25) and “reentry” (above Mach 25). Hypersonic speeds are accompanied, in general, by a small shock standoff distance. As the Mach number increases, the entropy layer of the air around the platform changes rapidly, and there are accompanying vortical flows. Also, a significant amount of aerodynamic heating causes the air around the platform to disassociate and ionize. From a flight test perspective, this matters because the plasma and the ionization interfere with the radio frequency (RF) channels. This interference reduces the telemetry links' reliability and backup techniques must be employed to guarantee the reception of acquired data. Additionally, the flight test instrumentation (FTI) package needs to perform optimally in and capture the higher acceleration, temperature, and vibration measurements that the hypersonic vehicle experiences.
Using current technologies, a single “entry level” vehicle has millions of electrical signals sent through dozens of modules, sensors and actuators, and those signals can be sent over the air, creating a telemetry data that can be used for several ends. One electrical device is set up to have diagnosis, in order to make maintenance feasible and support repair, plus giving improvement directions for specialists on new developments and specifications, but in several cases the diagnosis can only determine the mechanism of failure, but not the event that triggered that failure. Current evaluation method involves teardown, testing and knowledge from the involved specialized team, but this implies in recovering of failed parts, which in larger automakers with thousands of dealers/repair shops, reduces the sample for analyses when there is a systemic issue with one component. This specificity is usual in Propulsions systems, regarding electro-mechanical devices, and sensors, also in electrochemical devices, such as batteries and others, when a systemic issue appears, the teardown reveals its failure, but now why it failed. Based on that information and needs a methodology using big data mining and tools combined with available telemetry data in order to detect statistically main events or contributors/variables that triggers a failure event. That sort of methodology is helpful and more agile since it doesn’t depend on recovering of parts to give directions of which potential event may trigger a failure event, supporting in systemic/application comprehension of any component failure which uses electrical signals monitored within vehicle, and doesn’t depend on extraction of failed components, it can use and consider every single failed vehicle, for one specific component, as basis for analyses and identification of failure event, which will support in systemic correction/improvement and adjustment/improvement of specification for future and specific developments.
A wireless device called the UroMonitor enables accurate, noninvasive monitoring of bladder pressure in patients with overactive bladder. It is the first device to enable catheter-free telemetric ambulatory bladder pressure monitoring in humans. The UroMonitor was developed as a noninvasive approach for assessing function of the lower urinary tract, without the need for catheter placement. The UroMonitor is a small, flexible device — no more than 2 in. across — that is placed into the patient’s bladder. Once in place, the device wirelessly transmits bladder pressure data to a small radio receiver taped to the lower abdomen.
The challenge faced by flight software engineers at the Laboratory for Atmospheric and Space Physics (LASP) at the University of Colorado Boulder became evident when tasked with developing the onboard software for NASA's new Climate Absolute Radiance and Refractivity Observatory (CLARREO) Pathfinder Reflected Solar mission. The goal of measuring Earth-reflected sunlight with an accuracy of 0.3 percent (k=1), surpassing existing sensors by five to tenfold, from an instrument mounted beneath the International Space Station (ISS), produced a complex set of requirements. The avionics needed to balance multiple functions, including a high-rate control law, numerous hard real-time deadlines, interfaces with half a dozen external subsystems, and management of commands, telemetry and fault protection, all while capturing high-resolution science images at 15 frames per second. Ensuring uninterrupted operation within the unforgiving environment of low-Earth orbit necessitated the software run on a single Arm Cortex-M1 microprocessor instantiated at 50 megahertz (MHz) inside a radiation tolerant field-programmable gate array (FPGA) (Microchip RTG4), with only 4MB of RAM. The engineers quickly realized that existing flight software solutions at LASP were not equipped to handle this combination of hard real-time performance and heavy multitasking within such a constrained processing environment.
With the increased demand for electricity due to the rapid expansion of EV charging infrastructure, weather events, and a shift towards smaller, more environmentally responsible forms of renewable sources of energy, Microgrids are increasing in growth and popularity. The integration of real time communication between all PGSs (Power Generating Sources) and loadbanks has allowed the re-utilization of waste electricity. Pop-up Microgrids in PSPS events have become more popular and feasible in providing small to medium size transmission and distribution. Due to the differing characteristics of the PGSs, it is a challenge to efficiently engage the combined PGSs in harmony and have them share and carry the load of the microgrid with minimal ‘infighting.’ Different Power generating sources each have their own personality and unique ‘quirks.’ With loadbanks being able to perform various functions automatically by monitoring and responding to individual PGSs needs and demands, efficiency is improved and waste electricity is diverted to where it is required. The main useful functions of loadbanks mentioned in this paper represent individual desired functions. Maximizing the symbiosis in this microsystem by creating a common network and protocol environment between all components in the microgrid, adaptive automation is achieved. Refinements in translation of protocols into a standardization of SAE J1939 and CANOPEN from RS485 and MODBUS protocols contribute to the robustness of the framework including loadbanks. The Loadbank translates J1939 PGN messages on the CANBUS and reacts according to parameters from various ECM’s in the Power Generating Units, EGT, Voltage, Current and electrical load among others. Remote telemetry minimizes fault finding and common alarms are easily addressed remotely. Instituting EV charging and battery storage charging infrastructure on the backend of parasitic loadbanks in the microgrid environment maximizes otherwise wasted electricity whilst performing automated functions.
This paper proposes a new distributed/zonal control architecture composed of generic control modules. The new architecture increases the number of available input and outputs and improves reliability through redundancy, helping electric vehicles (EVs) meet the demand for advanced vehicle features and reliability. Each control module is equipped with analog and digital input pins, digital output pins, 2 CAN bus connections, power supplies from the 12V EV battery, a state-of-the-art DSP, and a wireless telemetry module for remote datalogging. A software architecture is developed to enable local signal processing from sensors, communication between modules in the distributed control architecture, and actuation of control objectives. Two CAN transceivers on two separate CAN busses are included for redundancy. Digi XBee Pro 900 HP module is used for long-distance wireless datalogging. TMS320F283874s microcontroller is used for its significant processing power and high number of peripherals. The controller units/boards were designed, fabricated and tested. Successful experimental validation was performed using the control module and emulator boards for the battery management system (BMS) and inverter. Multiple modules are connected and their ability to receive information from sensors, communicate with other boards, and control external hardware is tested. The prototype developed is intended for use in a Formula SAE EV.
As connected and automated vehicle technologies emerge and proliferate, lower frequency vehicle trajectory data is becoming more widely available. In some cases, entire fleets are streaming position, speed, and telemetry at sample rates of less than 10 seconds. This presents opportunities to apply powertrain simulators such as the National Renewable Energy Laboratory’s Future Automotive Systems Technology Simulator to model how advanced powertrain technologies would perform in the real world. However, connected vehicle data tends to be available at lower temporal frequencies than the 1-10 Hz trajectories that have typically been used for powertrain simulation. Higher frequency data, typically used for simulation, is costly to collect and store and therefore is often limited in density and geography. This paper explores the suitability of lower frequency, high availability, connected vehicle data for detailed powertrain simulation. A large data set of 1 Hz trajectories is used to quantify the accuracy loss when simulating energy consumption for conventional, hybrid, and battery electric powertrains using less than 1 Hz data. Techniques to upsample lower frequency drive cycle data in order to increase accuracy are also explored. Median energy consumption errors when simulating energy consumption for a 1/10 Hz trajectory are found to be 3-6% when compared to 1 Hz trajectories. Applying upsampling and interpolation techniques are shown to reduce the simulation errors by roughly 50%. The findings in this work can guide connected vehicle data collection specifications and processing techniques applied when using collected data for powertrain simulation.
Around the turn of this century, the automotive industry introduced a new type of technology to drive the gauges on a vehicle’s instrument cluster. The change was unannounced to the collision reconstruction world, but soon after, investigators observed a marked increase in crashed vehicles displaying frozen gauges at what often appeared to be correct readings. The new technology was the use of stepper motors which require power to return to the zero position. Hence if electrical power is lost, the gauges stop in position. There have been a number of previous papers covering the operation of the instruments and crash testing of cars and motorcycles to establish the ability of the instruments to withstand the forces on the instrument during a collision. This paper aims to compare the frozen instrument readings from real world collisions with the available EDR data from the crashed vehicles. With the assistance of the collision reconstruction community, a large dataset of 236 vehicles with frozen speedometer readings were compared with EDR and other corroborating methods. This paper reviews the current state of knowledge, compares the instrument readings of each of the 236 vehicles against the available EDR data or other corroborating method. It then assesses each case against the criteria proposed by Goddard and Price [3, 4] to assess if the cases that are being filtered out for selection, are accurate readings. It was found the existing criteria was an effective filter in removing the majority of cases capable of producing erroneous readings. However, it was found that some erroneous readings were present in low speed impacts. With the addition of a minimum speed criteria, the likelihood of an erroneous reading was greatly reduced. At recorded speedometer readings over 80 Km/h, the corroboration with the EDR speed had a standard deviation of 4%.
The logistics process in Brazil and the world represents a significant portion of the cost of manufactured products, either for export or import. The availability of technologies that make the logistic process more efficient directly affects the product’s transportation productivity and makes them more competitive. This paper presents a telemetry model of commercial vehicles integrated with harvest machines in agriculture operations, allowing accurate scheduling of loading and unloading processes at the field. In this study, we introduce a conceptual model of a technological matrix, where the shared topologies of vehicle information processing help predict failures, identification of wear of vehicle and machine’s components. The opportunity is demonstrated to collect data from agricultural machines and combine them with data extracted from trucks. The sharing of information on farm machinery and trucks in real-time establishes an essential change in crop management in the field.
This paper investigates the application of torque weighting to vibration dose value. This is done as a means to enhance correlation of perceived drive comfort directly to driver pedal commands while rejecting uncorrelated inputs. Current industry standards for vehicle comfort are formulated and described by ISO2631, which is a culmination of research with single or multi-axis vibration of narrow or broadband excitation. The standard is capable of estimating passenger comfort to vibrations, however, it only accounts for reaction vibrations to controlled inputs and not perceived vibration request vs. response vibration. Metrics that account for torque inputs and the vibration response create actionable estimates of dosage due to driver torque requests without uncorrelated inputs. This reduces the need for additional accelerometers and special compensating algorithms when road or track testing. The use case for the proposed modified metric is during the powertrain calibration process. Specifically, it can be used to evaluate driver commanded torque transients that cause torque reversal(s) of the drivertrain, e.g., coast to drive or drive to coast, through pedal tip-in or tip-out, respectively. Two body on frame, solid rear axle, full-size trucks featuring a twin turbocharged gasoline direct injected engines each paired with a ten-speed automatic transmission and four-wheel drive selectable transfer case each were utilized in the investigation to collect test data. Road testing was performed on both vehicles instrumented with accelerometers, telemetry torque meters and CAN signal data. Vehicle NVH data in combination with standard and modified dose metrics acquired in various gear states, powertrain calibration configurations and vehicle loading states is presented. The utilization of transient event based torque weighting will be shown to improve correlation of subjective driver related NVH measurements to an objective quantity of dose value for the purposes of adjusting powertrain calibration.
There is a need for rad-hard crystal stabilized clock sources with at least 300 krad of total ionizing dose (TID) immunity. A common solution has been to spot-shield a commercial off-the-shelf part or enclose it in a vault. Rad-hard clock sources are needed for main electronics boards (MEBs) and readout electronics that need to operate in hazardous space environments. Remote sensing and telemetry require that the readout circuits be co-located with the sensors, which can be separated by an arbitrary distance from the data processing electronics.
Rule-based systems seem natural for runtime verification (RV)/program monitoring. From a specification notation point of view, rule-based systems appear quite suitable for expressing the kind of properties the runtime verification community normally writes. Specifications written in a rule system have an operational flavor, which can be seen as a disadvantage or an advantage, depending on the viewpoint. The operational flavor makes specifications longer than in declarative temporal logic or regular expressions; however, they are natural to write. Once the core idea is mastered, writing rules is straightforward, like programming. More declarative specifications can be trickier to get right. This observation is similar to the observation that it may be easier to formulate a nontrivial property as a state machine than as a temporal logic formula or a regular expression.
Increased electrification of future heavy-duty engines and vehicles can enable many new technologies to improve efficiency. Electrified oil pumps are one such technology that provides the ability to reduce or turn off the piston oil cooling jets and simultaneously reduce the oil pump flow to account for the reduced flow rate required. This can reduce parasitic losses and improve overall engine efficiency. In order to study the potential impact of reduced oil cooling, a GT-Power engine model prediction of piston temperature was calibrated based on measured piston temperatures from a wireless telemetry system. A simulation was run in which the piston oil cooling was controlled to target a safe piston surface temperature and the resulting reduction in oil cooling was determined. With reduced oil cooling, engine BSFC improved by 0.2-0.8% compared to the baseline with full oil cooling, due to reduced heat transfer from the elevated piston temperatures. A GT-Drive vehicle model was used to evaluate the potential of this technology on a heavy-duty vehicle over various drive cycles. The parasitic losses of the engine oil pump were removed from the engine fuel efficiency map, and electric pump was added to the vehicle model. The pump flow rate was optimized to match the requirements of the engine at any given speed and load condition. The oil pump flow rate was further optimized based on the amount of piston oil cooling reduction possible at each operating point. The model was exercised over four drive cycles and vehicle fuel economy was observed. Results for the simulation showed that an electrified oil pump combined with reduced piston oil cooling provided an average fuel economy improvement of 2.7% over the baseline vehicle.
Space Dynamics Laboratory Utah State University Logan, UT
ABSTRACT This paper describes the use of neural networks to enhance simulations for subsequent training of anomaly-detection systems. Simulations can provide edge conditions for anomaly detection which may be sparse or non-existent in real-world data. Simulations suffer, however, by producing data that is “too clean” resulting in anomaly detection systems that cannot transition from simulated data to actual conditions. Our approach enhances simulations using neural networks trained on real-world data to create outputs that are more realistic and variable than traditional simulations. Citation: P.Feldman, “Training robust anomaly detection using ML-Enhanced simulations”, In Proceedings of the Ground Vehicle Systems Engineering and Technology Symposium (GVSETS), NDIA, Novi, MI, Aug. 11-13, 2020.
Several GoPro camera models contain Global Positioning System (GPS), accelerometer, and gyroscope instrumentation and are capable of measuring and recording position, velocity, acceleration, and inertial data. This study evaluates the accuracy of data obtained from GoPro cameras through a series of controlled tests. A test vehicle was instrumented with a Racelogic VBOX data acquisition unit as well as various generations of GoPro camera units equipped with GPS capability and driven on a road course. The raw data collected with the GoPro cameras and the translations of this data provided by the GoPro Quik desktop software application were compared to data collected with the validated VBOX data acquisition unit. The results demonstrated that position, velocity, and acceleration data recorded with GoPro cameras is consistent with VBOX data and is useful for applications related to accident reconstruction.
There is increasing demand for high-quality High Definition (HD) video for airborne applications such as Flight Test Instrumentation (FTI). Ideally, such new camera solutions can reduce the weight and difficultly of installing wiring, and enable data to be coherently combined with image data. Ethernet cameras can address these needs with built-in compression and multiple output streams. Additionally, as Ethernet-based networks have become an attractive choice for FTI applications, we see increased requirements for integrating Ethernet-based cameras with FTI data acquisition equipment, network recorders, and telemetry systems as this removes duplication of wiring and devices.
The effectiveness of using neural networks to predict rotor loads on the AW609 tilt-rotor is proven in this work. The main objective is to find a viable architecture for a neural network simple enough to be implemented in real time, with the aim to have a reliable prediction of rotor loads during telemetry monitoring sessions of flight test operations. The real time comparison of the loads predicted by the neural network with those measured by the aircraft instrumentation can provide immediate hints of incipient anomalies. A simple Feed Forward neural network has been tested, analyzing briefly the pros and cons of such a choice versus other possible architectures. The proposed neural network will estimate the bending loads (beam and chord) and the pitch link axial load, given the parameters that describe the aircraft trim point and how it is maneuvering. Instead of trying to estimate directly the time history of the loads, with all its associated dynamics, an approach based on a harmonic decomposition is here proposed. In particular, the signal is first decomposed in its harmonic components and various neural networks are trained efficiently to predict a single harmonic at a time. The complete time history is then reconstructed α-posteriori by combining all the signals predicted by the different neural networks.
Worldwide demand for low Earth orbit satellites is increasing at an unprecedented pace, driven by diverse needs such as faster and more affordable Internet access, and faster revisit rates with finer resolution for imaging data. The satellite payload instruments performing communications or imaging functions are becoming increasingly sophisticated and capable and require the collection of increasing amounts of telemetry data to ensure the safe and reliable operation of the satellite.
Items per page:
50
1 – 50 of 374