Driving Behavior Modelling with Graph Attention Networks

<div class="section abstract"> <div class="htmlview paragraph">This document provides preliminary<sup><span class="xref">1</span></sup> safety-relevant guidance for in-vehicle fallback test driver training and for on-road testing of vehicles being operated by prototype conditional, high, and full (Levels 3 to 5) ADS, as defined by SAE J3016. It does not include guidance for evaluating the performance of post-production ADS-equipped vehicles. Moreover, this guidance only addresses testing of ADS-operated vehicles as overseen by in-vehicle fallback test drivers (IFTD).</div> <div class="htmlview paragraph">These guidelines do not address:</div> <ul class="list disc"> <li class="list-item"> <div class="htmlview paragraph">Remote driving, including remote fallback test driving of prototype ADS-operated test vehicles in driverless operation. (Note: The term “remote fallback test driver” is included as a defined term herein and is intended to be addressed in a future iteration of this document. However, at this time, too little is published or known about this type of testing to provide even preliminary guidance.)</div></li> <li class="list-item"> <div class="htmlview paragraph">Testing of driver support features (i.e., Levels 1 and 2), which rely on a human driver to perform part of the dynamic driving task (DDT) and to supervise the driving automation feature’s performance in real time. (Refer to SAE J3016.)</div></li> <li class="list-item"> <div class="htmlview paragraph">Closed-course testing.</div></li> <li class="list-item"> <div class="htmlview paragraph">Simulation testing (except for training purposes).</div></li> <li class="list-item"> <div class="htmlview paragraph">Component-level testing.</div></li></ul> <div class="htmlview paragraph">These guidelines also do not address prototype vehicle and IFTD performance data collection and retention. The collection of data invokes various legal and risk management considerations that users of this document should nevertheless bear in mind, such as:</div> <ul class="list disc"> <li class="list-item"> <div class="htmlview paragraph">Maintaining auditable procedures and documentation.</div></li> <li class="list-item"> <div class="htmlview paragraph">Adhering to applicable privacy laws and principles.</div></li> <li class="list-item"> <div class="htmlview paragraph">Ensuring adequate data collection and recording integrity to support post-crash forensic analysis.</div></li></ul></div>

<div class="section abstract"><div class="htmlview paragraph">In traditional four-wheeled automobiles, the imbalance between the roll moment, which is the product of the centrifugal force during a turn acting on the center of gravity and the height of the center of gravity, and roll stiffness, which is the product of the left-right difference in tire vertical load and the tread width and commonly used among automotive suspension engineers, of the front and rear sections necessitates body torsional rigidity. However, there is a lack of specific cases and guidelines for constructing the body structure of three-wheeled PMVs (Personal Mobility Vehicles) with a tilting mechanism from the perspective of vehicle dynamics characteristics. In this paper, the basic considerations related to the dynamics of such three-wheeled PMVs are investigated.</div><div class="htmlview paragraph">We use the term “torsional rigidity” to refer to the stiffness as the torsional deformation of the body itself, and the term “roll stiffness” to refer to the moment that counteracts the roll moment during a turn and suppresses the vehicle’s roll in accordance with the terminology commonly used among automotive suspension engineers. In this paper, “torsional rigidity” and “roll stiffness” are clearly distinguished.</div><div class="htmlview paragraph">Through a systematic comparative analysis with conventional four-wheeled automobiles, this study theoretically formulates the concept of mechanical equilibrium for the body torsional rigidity of three-wheeled PMVs, resulting in fundamental guidelines governing their body structure.</div><div class="htmlview paragraph">The findings that the tilting function does not seem to impose torsional loads on the body during steady turning as analyzed through static analysis, shows that torsional rigidity around the longitudinal axis may be unnecessary. However, under actual dynamic conditions, a different scenario emerges. Dynamic simulations under real-world conditions demonstrate significant torsional loads on the body, emerging the necessity of torsional rigidity around the longitudinal axis.</div><div class="htmlview paragraph">In three-wheeled PMVs, body torsion directly influences dynamic vehicle response. Therefore, to accurately delineate the structural requirements for body design, quantitative evaluation of the necessary body torsional rigidity under dynamic conditions is essential. Future efforts will involve creating a dynamic simulation model that integrates considerations for body torsion based on these initial study outcomes. Ultimately, this research aims to bridge the gap between theoretical understanding and practical implementation, contributing to comprehensive guidelines for the body design of three-wheeled PMVs.</div></div>

<div class="htmlview paragraph">SANDEN CORPORATION (named SANKYO ELECTRIC CO., LTD. prior to September 1982) developed a new automobile air conditioning compressor, the "TR" type, based upon the spiral machine principle.</div> <div class="htmlview paragraph">Two spiral components have 2.8 turn involute spiral walls projected from discs. Both spiral components are housed in a cylindrical casing, with one in a fixed position and the other interleaved with the former at an angular relationship of 180°. The free moving component is driven by a crankshaft in an orbiting motion. The orbiting motion of the free moving component provides a continuous compression process in two pairs of compression pockets.</div> <div class="htmlview paragraph">The "TR" compressor has several unique features in its physical and mechanical design: (1) Specifically disposed and shaped spiral components to maximize the displacement within a small diameter casing, (2) A compliant crank mechanism utilizing the swing motion of an eccentric bush, (3) A perfect dynamic balancing mechanism with minimized radial forces, and (4) An anti-rotation coupling mechanism with a function of the orbiting thrust bearing.</div> <div class="htmlview paragraph">These basic features result in a number of advantages in application which are of interest to today's automotive design engineers: <ol class="list nostyle"> <li class="list-item"> <span class="li-label">(a)</span> <div class="htmlview paragraph">High refrigeration capacity from a compact, lightweight package achieved by combining the super high speed capability (up to 13,000 rpm) and a high volumetric efficiency (75% to 85%).</div></li> <li class="list-item"> <span class="li-label">(b)</span> <div class="htmlview paragraph">Extremely smooth operation with low torque requirement and inherently low torque variation.</div></li> <li class="list-item"> <span class="li-label">(c)</span> <div class="htmlview paragraph">Weight reduction of the compressor and its mounting means.</div></li> <li class="list-item"> <span class="li-label">(d)</span> <div class="htmlview paragraph">High performance, comfortability and drivability combined with low N.V.H. (Noise, Vibration, and Harshness) characteristics.</div></li></ol></div>

<div class="section abstract"><div class="htmlview paragraph">All internal combustion piston engines emit solid nanoparticles. Some are soot particles resulting from incomplete combustion of fuels, or lube oil. Some particles are metal compounds, most probably metal oxides. A major source of metal compound particles is engine abrasion. The lube oil transports these abraded particles into the combustion zone. There they are partially vaporized and ultrafine oxide particles formed through nucleation [<span class="xref">1</span>]. Other sources are the metallic additives to the lube oil, metallic additives in the fuel, and debris from the catalytic coatings in the exhaust-gas emission control devices. The formation process results in extremely fine particles, typically smaller than 50 nm. Thus they intrude through the alveolar membranes directly into the human organism. The consequent health risk necessitates a careful investigation of these emissions and effective curtailment.</div><div class="htmlview paragraph">Substantial information is available on diesel engine particulate emissions [<span class="xref">2</span>, <span class="xref">3</span>, <span class="xref">4</span>] but less for SI engines. Beside an example of metal oxide particles from a diesel engine, the present paper shows preliminary results of nanoparticles emissions of SI engines.</div><div class="htmlview paragraph">Four SI engines were investigated: two older and two newer engines, comprising two car engines and two motorbikes. The tests were done on standard transient driving cycles, and steady-state at constant 50 km/h and idling. Prior investigations observed, especially during idling, high concentrations of metal oxide particles [<span class="xref">2</span>]. All tests were done with particle samples collected from the CVS tunnel, during a long operating period, to have sufficient material for analyzing the composition. At the steady-state points, the particle size spectra were measured.</div><div class="htmlview paragraph">The results show that the older SI engines emit high concentrations of soot particles and metal oxide (= ash) particles. The size distribution is decisively bimodal for both soot and ash particles.</div><div class="htmlview paragraph">The newer engines' emission results are less uniform and the concentrations lower, as expected. Altogether, the concentrations of these metal oxide particles are so high, that more detailed investigations are recommended.</div></div>

<div class="section abstract"><div class="htmlview paragraph">In automobile design, a rear-engine layout mainly espoused for small entry-level cars and light commercial vehicles for three reasons - packaging, traction, and ease of manufacturing. The aim of this paper is to strategize cooling system of rear-mounted engine of a small gasoline car. Radiator and cooling fan packaged close to engine at rear of the car for simple packaging. Efficient thermal management ensured by robust overheat protection stratagem using EMS software.</div><div class="htmlview paragraph">DFSS, a disciplined problem prevention approach that helps in achieving the most optimum design solution and provides improved and cost effective quality products; is used to finalize an optimum design based on the analysis of the various tests carried out as per DOE [<span class="xref">1</span>]. This paper is about designing a distinctive cooling system of a car having rear-mounted engine with rear radiator but front mounted HVAC system [<span class="xref">2</span>].</div><div class="htmlview paragraph">Strategizing overheat protection for subject vehicle is crucial due to three critical reasons which tempt high probability of engine overheating. (1) The arrangement having HVAC system mounted at front requires lengthy coolant lines extending from engine (rear of the car) to the HVAC heater (front of the car). The underbody assembly of heater lines are prone to structural damage thus leading to overheating. (2) Lengthy coolant lines of HVAC heater system becomes critical as air prone to trap in the coolant system during coolant filling and de-aeration process leading to overheating. (3) In case of radiator fan failure, rear mounted engine compartment does not get enough surrounding air for cooling and will lead to overheating.</div><div class="htmlview paragraph">This final design provide compact packaging of coolant system, ensure better thermal management and maintain good aesthetics of the automobile. The serviceability of the engine i.e. filling of the coolant and process of de-aeration also simplified with the finalized design.</div><div class="htmlview paragraph">This paper also comprise a robust engine protection EMS (Engine Management System) function in case of overheat phenomenon. The strategy validated on fleet cars and it meets safety critical requirements to protect engine and critical components from damage. This novel FIVE LEVEL monitoring strategy ensures protection of vehicle and components in the event of cooling system failure.</div></div>

<div class="section abstract"><div class="htmlview paragraph">WITH the advent of low-pressure tires and the success of car noise reduction efforts by automotive engineers, tire thump has become a more noticeable problem. This paper discusses some recent research in this area.<span class="xref">*</span></div><div class="htmlview paragraph">A tire's susceptibility to thump can be predicted from the difference frequency between diametrical and circular modes of vibration. The author describes the testing and procedure uses to find the source of thump.</div><div class="htmlview paragraph">The way to minimize thump is to shape and construct the tire so that it does not permit coincidence of two natural frequencies with two revolution rate harmonics in the 20–35 mph speed range.</div></div>

<div class="htmlview paragraph">The feasibility of a simple-cycle gas turbine engine for automobiles is analyzed by means of comparison with Otto-cycle engines for automobiles. Applicable technology is reviewed with emphasis on centrifugal compressor technology, which has demonstrated pressure ratios of better than 10:1. The design point of a simple-cycle 150 hp automobile engine based on these components is presented, together with estimated torque and fuel-flow characteristics for the entire range of possible engine performance, derived from engine-cycle analysis programs.</div> <div class="htmlview paragraph">The results of a vehicle performance computer program designed to evaluate candidates for this application include time/speed/distance performance estimates for maximum acceleration as well as fuel economy and emissions estimates for pertinent driving cycles. Summaries of manufacturing cost estimates are presented for several candidate gas turbine engines.</div> <div class="htmlview paragraph">A speculative overall cost for both four and ten years of ownership is estimated based primarily on these fuel economy and manufacturing cost derivations.</div> <div class="htmlview paragraph">It is concluded that the simple-cycle gas turbine engine studied is sufficiently attractive as a low-pollution automobile engine for the post-1975 decade to warrant full-scale demonstration as soon as possible.</div>

<div class="section abstract"><div class="htmlview paragraph">The Future Automotive Systems Technology Simulator (FASTSim) is a high-level advanced vehicle powertrain systems analysis tool supported by the U.S. Department of Energy's Vehicle Technologies Office. FASTSim provides a quick and simple approach to compare powertrains and estimate the impact of technology improvements on light- and heavy-duty vehicle efficiency, performance, cost, and battery life. The input data for most light-duty vehicles can be automatically imported. Those inputs can be modified to represent variations of the vehicle or powertrain. The vehicle and its components are then simulated through speed-versus-time drive cycles. At each time step, FASTSim accounts for drag, acceleration, ascent, rolling resistance, each powertrain component's efficiency and power limits, and regenerative braking. Conventional vehicles, hybrid electric vehicles, plug-in hybrid electric vehicles, all-electric vehicles, compressed natural gas vehicles, and fuel cell vehicles are included. Powertrains with electric-traction drive can optionally be simulated using electric roadway technologies such as dynamic wireless power transfer. FASTSim also has an interface for running large batches of real-world drive cycles. FASTSim's calculation framework and balance among detail, accuracy, and speed enable it to simulate thousands of driven miles in minutes. The key components and vehicle outputs have been validated by comparing the model outputs to test data for many different vehicles to provide confidence in the results. A graphical user interface makes FASTSim easy and efficient to use. FASTSim is freely available for download from the National Renewable Energy Laboratory's website (see <a href="http://www.nrel.gov/fastsim" target="_blank">www.nrel.gov/fastsim</a>). <figure id="F1" class="figure"><div class="graphic-wrapper"><img class="article-figure figure" src="2015-01-0973_fig0001.jpg" alt="No Caption Available"/></div></figure></div></div>

<div class="htmlview paragraph">The use of graphical dynamic system simulation software is becoming more popular as automotive engineers strive to reduce the time to develop new control systems. The use of model-based control methods designed to meet future emission and diagnostic regulations has also increased the need for validated engine models. A previously validated, nonlinear, mean-torque predictive engine model<span class="xref">*</span> is converted to MATLAB / SIMULINK<span class="xref"><sup>†</sup></span> to illustrate the benefits of a graphical simulation environment. The model simulates a port-fuel-injected, spark-ignition engine and includes air, fuel and EGR dynamics in the intake manifold as well as the process delays inherent in a four-stroke cycle engine. The engine model can he used in five ways:</div> <div class="htmlview paragraph"> <ul class="list disc"> <li class="list-item"><div class="htmlview paragraph">As a nonreal-time engine model for testing engine control algorithms</div> </li> <li class="list-item"><div class="htmlview paragraph">As a real-time engine model for hardware-in-the-loop testing</div> </li> <li class="list-item"><div class="htmlview paragraph">As an embedded model within a control algorithm or observer</div> </li> <li class="list-item"><div class="htmlview paragraph">As a system model for evaluating engine sensor and actuator models</div> </li> <li class="list-item"><div class="htmlview paragraph">As a subsystem in a powertrain or vehicle dynamics model</div> </li> </ul> </div> <div class="htmlview paragraph">Although developed and validated for a specific engine, the model is generic enough to be used for a wide range of spark-ignition engines. Modular programming techniques reduce model complexity by dividing the engine and control system into hierarchical subsystems.</div>

<div class="htmlview paragraph"><i>The rapid trend in recent years toward higher octane fuels and high-performance engines has suggested a reappraisal of the ability of current laboratory knockrating methods to define the antiknock behavior of commercial gasolines in modern multicylinder passenger car engines. Road antiknock performance data have been obtained on 91 premium gasolines in five 1957 test cars and three cars equipped with experimental 11-to-1 compression ratio engines. In addition, 16 prototype gasolines were studied in the three high compression ratio cars.</i></div> <div class="htmlview paragraph"><i>The data were analyzed by establishing a linear relationship between the road ratings and the laboratory data to determine the relative importance of each variable in enhancing the accuracy of the prediction equation. Of the many fuel laboratory variables, combinations of Research and Motor octane number, fuel composition and volatility were investigated as predictors of road performance.</i></div> <div class="htmlview paragraph"><i>For any given combination the results were rather consistent from car to car, but varied greatly with car speed. Research octane number was found to be the best single predictor of low-speed road performance (2000 rpm), and Motor octane number the best single predictor of high-speed performance (3000 rpm and higher).</i></div> <div class="htmlview paragraph"><i>If two laboratory variables are used as predictors, a combination of Research and Motor octane numbers serves best for predicting low-speed road performance. At high speed, the combination of Motor octane number and percent aromatics was found to be the best two-variable predictor.</i></div> <div class="htmlview paragraph"><i>Low-speed correlations based on Research and Motor octane number were not improved when olefin content was included as a third predicting variable. At high speed, however, the use of a hydrocarbon-type variable in addition to Research and Motor octane numbers resulted in a significant reduction of the standard error of prediction.</i></div> <div class="htmlview paragraph"><i>The use of more than three variables as predictors did not improve the accuracy of prediction at low speed and improved the accuracy at high speeds to a degree which is not considered to be of practical value.</i></div> <div class="htmlview paragraph"><i>An increase in Research octane number was found to be of more importance than an increase in Motor octane number in improving low-speed road ratings. At high speeds, however, the contribution of Motor octane number was decidedly greater than that of Research octane number.</i></div> <div class="htmlview paragraph"><i>When Research and Motor octane numbers were held constant, an increase in aromatic content or a decrease in olefin content improved the high-speed road rating of a fuel</i></div>

<div class="section abstract"><div class="htmlview paragraph">Strict emission regulations and the need of higher efficiency of future dual fuel engines require an optimized combustion process. For getting a better understanding of the in-cylinder combustion process optical investigations represent a powerful tool. For medium speed dual fuel engines, optical investigations are pretty rare respectively not available. Especially the avoiding of knock events within the combustion process is a key development topic to realize high engine load and high engine efficiency.</div><div class="htmlview paragraph">For the investigations a fully flexible dual fuel test engine was used. The engine is operated with a natural gas / air cylinder charge which is ignited by a small micro pilot diesel injection within the gas mode.</div><div class="htmlview paragraph">Beside standard development measurement techniques like cylinder pressure measurement and exhaust gas emission measurement, optical measurement techniques were used to gain information about the in-cylinder combustion process of the pilot fuel combustion as well as the combustion of the natural gas / air cylinder charge. It is desirable to realize an entire visualization of the combustion within the combustion chamber under real engine conditions at full load. This means the need for a minimal invasive measurement technique with small modifications of the combustion chamber design and surface temperatures.</div><div class="htmlview paragraph">Due to these requirements a tomographic, fiber based measurement technique was used synchronal to a high speed visualization of the in-cylinder OH* chemiluminescence. The fiber based measurement technique uses 128 optical fibers in which every single fiber measures the luminescence of a special observation volume. The raw data of the single fibers is used to calculate, based on a tomographic algorithm, the in-cylinder light emission intensity distribution to visualize the combustion in a horizontal cutting plane over the whole bore diameter. To complete the information of the combustion by the third dimension, a synchronized, intensified high speed camera was used to record the OH* chemiluminescence. Both measurement techniques in combination give a far-reaching image of the in-cylinder combustion process.</div><div class="htmlview paragraph">To increase the understanding of the in-cylinder processes and to optimize the combustion process, typical combustion relevant parameters like pilot injection nozzle, pilot injection timing, pilot injection pressure and excess air ratio were varied. In this publication the basic combustion mechanisms are described for two different injection timings at full load. In addition the isolated pilot fuel combustion in pure air for an injection timing which gives a transient point between partly premixed and strongly premixed combustion is shown.</div><div class="htmlview paragraph">The measurement results show an ignition process which is strongly influenced by the reaction kinetic mechanisms. Especially for early injection timings a long ignition delay was measured which result in a well premixed pilot fuel within the natural gas / air mixture. Moreover, the flame propagation speed was calculated and a reaction kinetically controlled post flame combustion process was observed.</div><div class="htmlview paragraph">In addition a different mechanism for in-cylinder pressure oscillation, which usually is detected as engine knock, was shown which differs from the generally known mechanism in spark plug ignited passenger car engines.</div></div>

<div class="section abstract"><div class="htmlview paragraph">Autonomous vehicles have the potential to provide mobility to individuals who experience transportation disadvantages due to the inability to drive as a result of physical, cognitive or visual limitations/impairments as well as able-bodied individuals with no/limited desire to drive. Individuals who do not have easy access to transportation have social, academic, health, and career disadvantages in comparison to their peers. Fully autonomous vehicles have the potential to offer mobility solutions to these individuals.</div><div class="htmlview paragraph">A user-centered design approach was utilized by a multidisciplinary team of engineers, human factors specialists, and designers to develop future vehicle features for a broad range of users. In-depth interviews and day-in-the-life experiences provided the team with an understanding of the needs and constraints of very different types of users including:<ul class="list disc"><li class="list-item"><div class="htmlview paragraph">a millennial with a visual impairment,</div></li><li class="list-item"><div class="htmlview paragraph">a professional in a major metropolitan area who does not want to drive yet needs to transport influential clients to meetings throughout the city,</div></li><li class="list-item"><div class="htmlview paragraph">a full-time working mom with 3 kids under the age of 12, and.</div></li><li class="list-item"><div class="htmlview paragraph">a designer who is a quadriplegic who lives as independently as possible while using a large motorized wheelchair.</div></li></ul></div><div class="htmlview paragraph">User preferences, needs, and wants are discussed for each user. Future autonomous vehicles will have vastly different metrics for user satisfaction than what exists today.</div></div>

<div class="section abstract"><div class="htmlview paragraph">Android Automotive OS (AAOS) has been gaining popularity in recent years, with several OEMs across the world already deploying it or planning to in the near future. Besides the benefit of a well-known, customizable and secure operating system for OEMs, AAOS allows third-party app developers to offer their apps on vehicles of several manufacturers at the same time. Currently, there are 55 apps for AAOS that can be categorized as media, navigation or point-of-interest apps. Specifically the latter two categories allow the third-parties to collect certain sensor data directly from the vehicle. Furthermore, the latest version of AAOS also allows the OEM to configure and collect In-Vehicle Infotainment (IVI) and vehicle data (called OEM telemetry).</div><div class="htmlview paragraph">However, increasing connectivity and integration with the in-vehicle network comes at the expense of user privacy. Previous works have shown that vehicular sensor data often contains personally identifiable information (PII). New privacy regulations around the world mandate that the collection and processing of this data has to be clearly communicated with the user of the vehicle who reserves the right to approve or deny. In this paper, the existing AAOS apps are manually analyzed for the user data they collect and share. Of particular interest is the consistency of the declared app permissions with developers’ privacy policies since discrepancies can suggest compliance issues. Our study results show that over 78% of analyzed apps do not mention all dangerous permissions in their privacy policies.</div></div>

<div class="htmlview paragraph"><b>AUTOMOBILE engines, Mr. Rabezzana declares, have reached a point where it is very difficult, by a single improvement, to render any decidedly better performance, adding that, in his opinion, any further worthwhile improvements in engine economy on a conventional engine will be made only by several well-directed moves concerning small details.</b></div> <div class="htmlview paragraph"><b>He points out that improvements in part-load economy can be obtained by using a homogeneous mixture and efficient ignition.</b></div> <div class="htmlview paragraph"><b>Multicylinder engines are the ones most affected by low economy at light loads, he maintains, and lists the following as the main disturbing factors inherent in the majority of them: non-uniform mixture distribution among cylinders; low mixture speed in the manifold at light loads; poor turbulence in the cylinder-head; stratification of the mixture; dilution of the mixture due to the valve timing overlapping the high exhaust back-pressure; poor scavenging around the spark-plug gap position; and improper mixture at the spark-plug gap.</b></div>

<div class="htmlview paragraph">Very few data seem to be available on the frictional losses in automobile engines caused by the failure of the oil to perform its function as a lubricant. The researches of the Lubrication Inquiry Committee in England indicate that the friction of a flooded bearing is proportional to the speed of the engine, the area of the bearing and the viscosity of the lubricant and is independent of the pressure and of the materials of which the opposing surfaces are composed.</div> <div class="htmlview paragraph">The principal sources of friction in an engine are the crankshaft, the camshaft and the connecting-rod bearings, which rotate; the pistons and the valves, which slide; and the auxiliaries, such as the generator, the pump and the distributor. Experiments by the author led to the conclusion that a variation in the viscosity of the lubricant can easily account for reduced mechanical efficiency and consequently for the increased consumption of gasoline, and that the friction-loss is proportional to the cubic capacity of the engine, whereas the brake mean effective pressure for the same horsepower is inversely proportional to the capacity. Tables are given showing the variation of the friction mean effective pressure at various speeds and at various temperatures of the water and the oil, the fuel consumption of the engine at part throttle and at varying oil and water temperatures and a method of calculating the reduction of the gasoline consumption, assuming a constant indicated thermal efficiency. The identity of the results, as determined from the friction curves and from actual tests, is said to suggest that the thermal efficiency is affected but little by the water-jacket temperatures and that the saving on account of the thermostats and the radiator shutters is caused almost entirely by the reduction in the viscosity of the oil. Because of thermostatic control the water temperature rises very rapidly to 120 deg. fahr. in short runs and to about 170 deg. fahr. after 15 or 20 min., but the oil temperature, even for periods of 30 min. of city driving, does not rise more than 40 deg. fahr. above atmospheric temperature, and after prolonged fast driving the rise amounts to about 90 deg. fahr. On the assumption that the temperature-difference between the oil and the atmosphere is independent of the atmospheric temperature, it is shown that to obtain a mean oil and water temperature of 170 deg. fahr., the condition of maximum efficiency, with water at 170 deg. fahr., the atmospheric temperature must be about 80 deg. fahr. if the car is driven at approximately 40 m.p.h., while for ordinary city driving the conditions of maximum efficiency are not realized unless the atmospheric temperature is about 130 deg. fahr. The suggested use of thin oil might not avail because in winter very severe conditions of load and speed are frequently encountered and a slight difference in the viscosity might be of great importance.</div> <div class="htmlview paragraph">In making a study of methods of keeping dirt out of an automobile-engine crankcase, it was found that certain oils that worked well under normal conditions became too viscous in cold weather. This was remedied by making the troughlike conduit to the pump suction of wire gauze through which the oil when sufficiently hot drops into the main body of oil in the oil-pan. The property of oiliness increases in importance as the bearing surfaces more nearly touch each other, and especially when the engine is starting from rest.</div>