CMHT Research |
This page contains information on the main areas of our research. Please feel free to browse and contact us if you have any questions or comments.
Since the early part of the last century automotive powertrain configurations have remained relatively constant with either spark ignited (gasoline) or compression ignited (diesel) engines coupled to a transmission and chassis. With the recent interest in hybridization of vehicle powertrains to reduce fuel consumption, there are many configurations of automotive powertrains being proposed and developed. These require the use of technology that has not had the benefit of 100 years of automotive use and refinement. This poses a challenge for today’s automotive powertrain engineers, designers, OEMs and suppliers. These new technologies must be tested and validated and integrated into complete powertrain systems yet no standard testing methods have been established by the industry. In addition many of these technologies dramatically increase the complexity of the vehicle control computers, new and additional techniques for development testing and in-use fault detection are required. McMaster is developing state of the art facilities for testing the components used in hybrid vehicles, including batteries, power electronics, hybrid transmissions, electric motors, internal combustion engines and complete vehicles. This test equipment combined with continued research in the areas of testing, test optimization, integration testing, fault detection and prognostics will enable McMaster to participate at the forefront of these exciting times in the automotive industry.
The ability to detect and diagnose faults is essential for the safe and reliable control of mechanical and electrical systems. In the presence of a fault, the system behaviour may become unpredictable, resulting in a loss of control which can cause unwanted downtime as well as damage to the system. There are two main types of methods to detect and diagnose faults: signal-based and model-based. Signal-based fault detection methods typically use thresholds to extract information from available measurements. This information is then used to determine if a fault is present. Model-based methods, as the name suggests, makes use of faults which can be modeled, typically through system identification. This type of fault detection and diagnosis is popular when well-defined models can be created and utilized. A number of fault detection and identification (FDI) methods are being implemented and developed by our research group. For example, one area of research involves developing a new neural network strategy to be used determining and predicting the presence of faults in engines. Another area of research involves the development of an interacting multiple-model approach, in an effort to correctly identify the presence of faults in an electrohydrostatic actuator (EHA).
Estimation theory plays an important role in a variety of fields – ranging from astronomy to commerce, and biochemistry to mechanics. It involves determining a value of some parameter of interest, typically by extraction from noisy observations. In engineering, one is usually concerned with the system states of a mechanical or electrical system. Quite often, the states are representative of the dynamics of the system. For example, one may be interested in the position, velocity, and acceleration of an actuator piston. Knowledge of these states is important for the successful control of the actuating device. The states are observed or measured by the use of sensors in the environment. However, measurements typically contain unwanted information such as system and measurement noise. It is the role of a filter to extract the useful information from the measurements, while minimizing the effects of noise and other unwanted disturbances. The smooth variable structure filter (SVSF) is a relatively new (2007) estimation strategy based on sliding mode theory, and has been shown to be robust to modeling uncertainties. It is a predictor-corrector method that makes use of an existence subspace and of a smoothing boundary layer to keep the estimates bounded within a region of the true state trajectory. This creates an inherently stable estimation process. This area of research involves advancing the development of the filter, and working on implementations (simulated and experimental) of the SVSF.
Automotive tracking systems are integral part of any autonomous driving system. A generic automotive tracking system performs several functions. The environment is perceived using sensors such as LiDAR, radar, vision sensor, ultrasonic sensor, and infrared camera. The acquired information is processed for detection and classification of objects of interest. Artificial intelligence and machine learning techniques are among pioneering approaches to tackle this issue. The noisy information about the objects are later employed to estimate the quantities of interest such as position, direction, velocity, acceleration, shape, size, etc. for each object. This information is essential for decision making process for autonomous driving functionalities. In CMHT we have developed our in-house car detection and tracking technology, which has led to a number of publications, and an experimental setup for on-road driving, data collection, and real-time detection and tracking. We aim to use information from different sensors and sensor fusion methods to extend the scope of the project.
Mechatronics is a multidisciplinary subject, involving mechanical, electrical, computer, control, and systems design engineering. In our research group, mechatronics engineering is used to unite the principles of mechanics, electronics, and computing to generate a simpler, more economical and reliable system. Recently, a novel electrohydrostatic actuator (EHA) prototype was created by combining a computer controlled servomotor with a hydraulic actuator circuit. An EHA is an emerging type of actuator typically used in the aerospace industry, and are self-contained units comprised of their own pump, hydraulic circuit, and actuating cylinder. The main components of an EHA include a variable speed motor, an external gear pump, an accumulator, inner circuitry check valves, a double-rod double-acting cylinder, and a bi-directional pressure relief mechanism. The objective of the design was to provide an efficient and accurate method of actuating heavy loads. A variety of control strategies were studied and implemented on the system in an effort to enhance the system performance. For example, strategies such as fuzzy logic, PID, feedforward, and sliding mode control were implemented. The results of this research have led to a number of conference and journal publications. |
Home | Research | Members | Publications | Contacts| Join Us