Search Results

This paper gives an overview of todays machine vision technology with specific emphasis on microcomputer-based image processing and its potential as a low cost machine vision system. Topics discussed are general hardware requirements, image enhancement and segmentation techniques for binary and gray level images, two dimensional shape analysis, and additional sensors to supplement two dimensional image information.

Machine vision technology is a tool being utilized in the new GMT-400 pickup truck Body Shops for process monitoring and control. These real-time Machine Vision Dimensional Gauging systems monitor 100% of the production's critical build features such as door and window openings, hinge locations, and fender mounting brackets, Traditional gauges typically can provide data on only a small sample of production −1% or less. Correlating the machine vision systems to master gauges allows accurate data to be collected on every job as it is being built. This complete dimensional control data provides information for process monitoring as well as a means to detect tooling adjustment requirements and the ability to detect build problems, even if they occur intermittently. Several methods of performing this correlation have been investigated, with the goal being to define a correlation procedure that works well in the plant environment.

Inlet valve deposits in gasoline engines have a significant effect on engine operation with particular reference to cold starting and driveability. Present methods of quantifying these deposits by weighing them or rating them with the aid of a visual rating scale are recognized as not being reliable indices of the detrimental effect of these deposits. A valve deposit quantification system was developed that relied on the use of machine vision. Algorithms were formulated to track the silhouetted edge profile of a backlit valve from which a valve volume was determined. The valve deposit volume was calculated as the difference in volume between the valve in its clean and coked states. The system was able to detect a minimum coke deposit level of 0.06g at the 95% confidence limit, the accuracy being based on the correlation between the volume as determined by the vision system and the mass of the deposit.

Machine vision systems are becoming an important part of the automotive manufacturing process with applications ranging from inspection through to process monitoring and control. The technology is indeed becoming vital to maintaining and enhancing the quality associated with each component from the smallest assembly to the entire vehicle body. This paper will examine two machine vision applications in the automotive industry. These installations have been recently developed to satisfy the needs of ensuring that the engine assembly process guarantees the highest standards of product quality. The first application discussed in this paper, describes how machine vision is being used to verify the ‘K’ Series engine valve timing gear, prior to undertaking a series of computer controlled automatic tests. The second case study describes an alpha-numeric character recognition system which is central to the selective assembly of bearing shells of the ‘K’ Series engine.

The effectiveness of statistical techniques in the management of manufacturing processes is reviewed, and difficulties in applying these techniques in the automotive stamping and assembly environment are discussed. The use of machine vision measurement to overcome these difficulties is described and examples of functioning installations are shown. The problem encountered in evaluating such systems in terms of quality improvement is explained, costs of product specification conformance and nonconformance are defined, and quality costs for U.S. and Japanese industry are compared. Reduced nonconformance cost is identified as the probable explanation for the Japanese advantage in the cost of quality comparison, and Japanese use of Taguchi's loss function is proposed as one of the mechanisms by which this has been accomplished.

This paper will describe two Machine Vision applications at the John Deere Plow and Planter Works. The Vision Inspection System utilizes three cameras mounted on a three axis Cartesian robot to perform three-dimensional measurements on a variety of formed sheet-metal parts. The system also maintains average & range statistical control charts on all measured dimensions. The Vision Welding System utilizes four stationary cameras and four stationary LASER units to locate the ends of weld seams on a variety of similar parts ranging in length from 990 mm to 2600 mm (39 inches to 102 inches). These end points are then used to offset the weld path which is then welded by a six axis articulated robot.

An experimental machine vision-based patch-sprayer was developed. This sprayer was primarily designed to do real-time weed density estimation and variable herbicide application rate control. However, the sprayer also had the capability to do high-resolution weed mapping if proper mapping techniques were integrated. Two weed mapping methods were developed. One was a GPS signal based off-line weed mapping; another one was radar distance measurement-based on-line weed mapping. The high-resolution weed maps provided evidence to further support the patch-spraying concept. Randomly sampled field images were processed with different nozzle control zone sizes and thresholding methods to simulate sprayer performance. Fundamental system design strategies regarding these two factors were obtained through simulation. System design techniques, including system construction, weed sensing and crop-row detection algorithms were reported.

Conventional methods of physicochemical models require various experts and a high measurement demand to achieve the required model accuracy. With an additional request for faster development time for diagnostic algorithms, this method has reached the limits of economic feasibility. Machine learning algorithms are getting more popular in order to achieve a high model accuracy with an appropriate economical effort and allow to describe complex problems using statistical methods. An important point is the independence from other modelled variables and the exclusive use of sensor data and actuator settings. The concept has already been successfully proven in the field of modelling for exhaust gas aftertreatment sensors. An engine-out nitrogen oxide (NOX) emission sensor model based on polynomial regression was developed, trained, and transferred onto a conventional automotive electronic control unit (ECU) and also proves real-time capability.

Accurate determination of driveshaft torque is desired for robust control, calibration, and diagnosis of propulsion system behaviors. The real-time knowledge of driveshaft torque is also valuable for vehicle motion controls. However, online identification of driveshaft torque is difficult during transient drive conditions because of its coupling with vehicle mass, road grade, and drive resistance as well as the presence of numerous noise factors. A physical torque sensor such as a strain-gauge or magneto-elastic type is considered impractical for volume production vehicles because of packaging requirements, unit cost, and manufacturing investment. This paper describes a novel online method, referred to as Virtual Torque Sensor (VTS), for estimating driveshaft torque based on Machine-Learning (ML) approach. VTS maps a signal from Inertial Measurement Unit (IMU) and vehicle speed to driveshaft torque.

Predicting airbag deployment geometries is an important task for airbag and vehicle designers to meet safety standards based on biomechanical injury risk functions. This prediction is also an extraordinarily complex problem given the number of disciplines and their interactions. State-of-the-art airbag deployment geometry simulations (including time history) entail large, computationally expensive numerical methods such as finite element analysis (FEA) and computational fluid dynamics (CFD), among others. This complexity results in exceptionally large simulation times, making thorough exploration of the design space prohibitive.

Electric Vehicles are subject to effects that lead to more or less rapid degradation of functions. This can cause hazards for the drivers and uninvolved road participants. For this reason, the must be detected and mitigated, to maintain the vehicle function even in critical situations until a safe operating mode can be established. This publication presents an intelligent digital twin, located in the edge and connected with an electric vehicle via 5G. That can improve the operation of electrified vehicles by enabling the online detection of abnormal situations in the electrified powertrain and vehicle dynamics. Its core component is the fault detection system, which is implemented based on a 1-Nearest Neighbor algorithm. It is initially trained on synthetic data, generated in CarMaker for real-world powertrain issues such as demagnetization and open-/short-switch failures, using detailed mathematical models.

Encapsulation of electric powertrains is a booming topic with the electrification of vehicles. It is an efficient way of reducing noise radiated by the machines even in later stages of the design and without altering the electromagnetic performance. However, it is still difficult to define the best possible treatment. The locations, thicknesses and material compositions need to be optimized within given constraints to reach maximum noise reduction while keeping added mass and cost at minimum. In this paper, a methodology to design the encapsulation based on numerical vibro-acoustic simulations is presented. In a first step, the covered areas are identified through post-processing of a finite element acoustic radiation model of the bare powertrain. In a second step, a design of experiment is performed to assess the influence of various cover parameters on the acoustic radiation results.

The use of standardized “bar-code” labels on all new vehicles would help improve the effectiveness of vehicle emission inspection programs. The label could contain information on the vehicle and its emission control system. This information could be used by automated emission testing equipment to select test standards and identify the types of emission control equipment present on the vehicle being tested. More accurate inspections result. The machine-readable labels could also be used by compatible equipment to provide servicing specifications which could result in more effective and efficient repairs.

New manufacturing technologies such as high speed machining (HSM) are being developed to produce high quality aerospace components. While our developing understanding of machining dynamics is enabling precise control of cutting tools to provide for high dimensional accuracy, residual stresses present in aluminum mill products can compromise the ability to machine dimensionally accurate components from these stock materials. The advantages of precise tool control can be lost if the metal being cut moves during machining. And, even a perfectly machined part that distorts when it is released from the machine bed will cause problems upon assembly. Thus, ensuring the quality of the mill product becomes an enabling technology for advanced manufacturing approaches such as HSM.

The Russians have developed a wide range of equipment for the forest industry. This report provides a profile of this general equipment line. The author concludes that the Russian power saws, track skidders and falling machines, trucks, and lower landings offer much in innovative design to American industry. On the other hand, their winchs, spars, and rubber-tired skidders need redesign.

Cost data are presented which provide justification for making systematic studies of all types of machining processes in order to select the most applicable tool materials. Full tool life tests are suggested for obtaining reliable machining data. Specific tool life test data are supplied for the drilling, reaming, and tapping of 4340 steel at 42 Rc, for the face milling of 32510 and 60003 grades of malleable iron, and for the comparative evaluation of T-15, and ultrahard high speed steels, C-8, titanium carbide base, and ceramic cutting tools. Sources of machining information are referenced along with a description of the Air Force Machinability Data Center.

Grey iron (G3000) is a class of iron that is used to manufacture a wide variety of components throughout the world. More than 32 million tons were poured in 1996 (1). The machinability of cast iron at various times is difficult and often cannot readily be linked to the manufacturing or casting processes. This recurring machinability problem coupled with an inability to positively identify its cause has been very costly. A closer look at the microstructural differences in castings revealed that there is a qualitative difference in the coarseness of the pearlite between parts that machine well and those that were difficult to machine in a production setting.

The paper discusses two methods to implement error compensation framework for NC machining. In the first case a novel real-time co-interpolator is utilized that has been developed and demonstrated on a machine tool equipped with an open architecture controller (OAC). The second solution features near-time reprocessing of the NC programs, that is more suitable for existing machining systems. A P-type, iterative learning control (ILC) algorithm is also presented for calculating the error compensation values. The paper concludes with the results of machining tests, showing the effectiveness of the error compensation methodologies.