Guanglie Zhang

Reconfigurable acceleration for Monte Carlo based financial simulation

This paper describes a novel hardware accelerator for Monte Carlo (MC) simulation, and illustrates its implementation in field programmable gate array (FPGA) technology for speeding up financial applications. Our accelerator is based on a generic architecture, which combines speed and flexibility by integrating a pipelined MC core with an on-chip instruction processor. We develop a generic number system representation for determining the choice of number representation that meets numerical precision requirements. Our approach is then used in a complex financial engineering application, involving the Brace, Gatarek and Musiela (BGM) interest rate model for pricing derivatives. We address, in our BGM model, several challenges including the generation of Gaussian distributed random numbers and pipelining of the MC simulation. Our BGM application, based on an off-the-shelf system with a Xilinx XC2VP30 device at 50 MHz, is over 25 times faster than software running on a 1.5 GHz, Intel Pentium machine

A hardware Gaussian noise generator using the Wallace method

We describe a hardware Gaussian noise generator based on the Wallace method used for a hardware simulation system. Our noise generator accurately models a true Gaussian probability density function even at high /spl sigma/ values. We evaluate its properties using: 1) several different statistical tests, including the chi-square test and the Anderson-Darling test and 2) an application for decoding of low-density parity-check (LDPC) codes. Our design is implemented on a Xilinx Virtex-II XC2V4000-6 field-programmable gate array (FPGA) at 155 MHz; it takes up 3% of the device and produces 155 million samples per second, which is three times faster than a 2.6-GHz Pentium-IV PC. Another implementation on a Xilinx Spartan-III XC3S200E-5 FPGA at 106 MHz is two times faster than the software version. Further improvement in performance can be obtained by concurrent execution: 20 parallel instances of the noise generator on an XC2V4000-6 FPGA at 115 MHz can run 51 times faster than software on a 2.6-GHz Pentium-IV PC.

Ziggurat-based hardware Gaussian random number generator

An architecture and implementation of a high performance Gaussian random number generator (GRNG) is described. The GRNG uses the Ziggurat algorithm which divides the area under the probability density function into three regions (rectangular, wedge and tail). The rejection method is then used and this amounts to determining whether a random point falls into one of the three regions. The vast majority of points lie in the rectangular region and are accepted to directly produce a random variate. For the nonrectangular regions, which occur 1.5% of the time, the exponential or logarithm functions must be computed and an iterative fixed point operation unit is used. Computation of the rectangular region is heavily pipelined and a buffering scheme is used to allow the processing of rectangular regions to continue to operate in parallel with evaluation of the wedge and tail computation. The resulting system can generate 169 million normally distributed random numbers per second on a Xilinx XC2VP3O-6 device.

2D Human Gesture Tracking and Recognition by the Fusion of MEMS Inertial and Vision Sensors

In this paper, we present an algorithm for hand gesture tracking and recognition based on the integration of a custom-built microelectromechanical systems (MEMS)-based inertial sensor (or measurement unit) and a low resolution imaging (i.e., vision) sensor. We discuss the 2-D gesture recognition and tracking results here, but the algorithm can be extended to 3-D motion tracking and gesture recognition in the future. Essentially, this paper shows that inertial data sampled at 100 Hz and vision data at 5 frames/s could be fused by an extended Kalman filter, and used for accurate human hand gesture recognition and tracking. Since an inertial sensor is better at tracking rapid movements, while a vision sensor is more stable and accurate for tracking slow movements, a novel adaptive algorithm has been developed to adjust measurement noise covariance according to the measured accelerations and the angular rotation rates. The experimental results verify that the proposed method is capable of reducing the velocity error and position drift in an MEMS-based inertial sensor when aided by the vision sensor. Compensating for the time delay due to the visual data processing cycles, a moving average filter is applied to remove the high frequency noise and propagate the inertial signals. The reconstructed trajectories of the first 10 Arabic numerals are further recognized using dynamic time warping with a direct cosine transform for feature extraction, resulting in an accuracy of 92.3% and individual numeral recognition within 100 ms.

Towards an ubiquitous wireless digital writing instrument using MEMS motion sensing technology

A micro inertial measurement unit (muIMU) which is based on MEMS accelerometers and gyro sensors is developed for real-time recognition of human hand motions, which when combined with appropriate filtering and transformation algorithms, becomes a digital writing system that can be used to record handwriting on any surface. The overall size of our muIMU is less than 25 mm times 70 mm times 20 mm, including the micro sensors, processor, and wireless interface components. We present our progress on using this muIMU based on Kalman filtering algorithm to filter the noise of sensors, which has allowed the system to successfully transform hand motions into recognizable and recordable English characters. Our goal is to implement this system to a digital hand-writing system that can interface with PC and mobile computing devices

An Attitude Compensation Technique for a MEMS Motion Sensor Based Digital Writing Instrument

A MAG-muIMU which is based on MEMS gyroscopes, accelerometers, and magnetometers is developed for real-time estimation of human hand motions. Appropriate filtering, transformation and sensor fusion techniques are combined in the ubiquitous digital writing instrument to record handwriting on any surface. In this paper, we discuss the design of an extended Kalman filter based on MAG-muIMU (micro inertial measurement unit with magnetometers) for real-time attitude tracking. The filter utilizes the gyroscope propagation for transient updates and correction by reference field sensors, such as gravity sensors, magnetometers or star trackers. A process model is derived to separate sensor bias and to minimize wideband noise. The attitude calculation is based on quaternion which, when compared to Euler angles, has no singularity problem. Testing with synthetic data and actual sensor data proved the filter will converge and accurately track the attitude of a rigid body. Our goal is to implement this algorithm for motion recognition of a 3D ubiquitous digital pen.

Development of a Human Airbag System for Fall Protection Using MEMS Motion Sensing Technology

This paper describes the development of a human airbag system which is designed to reduce the impact force from falls. A micro inertial measurement unit (muIMU), based on MEMS accelerometers and gyro sensors is developed as the motion sensing part of the system. A recognition algorithm is used for real-time fall determination. With the algorithm, a microcontroller integrated with the muIMU can discriminate falling-down motion from normal human motions and trigger an airbag system when a fall occurs. Our airbag system is designed to have fast response with moderate input pressure, i.e., the experimental response time is less than 0.3 second under 0.4 MPa. In addition, we present our progress on using support vector machine (SVM) training together with the muIMU to better distinguish falling and normal motions. Experimental results show that selected eigenvector sets generated from 200 experimental data sets can be accurately separated into falling and other motions

Handwriting tracking based on coupled μIMU/electromagnetic resonance motion detection

We have recently developed a ubiquitous digital writing instrument system based on a micro inertial measurement unit (mulMU), which consists of MEMS (micro- electro-mechanical system), accelerometers and gyroscopes, to compute the position of a marker through double integration of the acceleration measured, so as to real-time record and recognize human handwriting motion in a large writing area, i.e., a large whiteboard or screen. Owing to the random errors that exist in the MEMS sensors, the accuracy of the position estimate degrades with time. Although Kalman filtering algorithm provides a good navigation tracking solution, its accuracy depends on the amount of position information given about the target. In vehicles, the global positioning system (GPS) can be used to augment an IMU with absolute position information and improve its tracking accuracy. However, due to indoor-usage and a higher accuracy requirement, the GPS is not suitable for updating a mulMU used for hand-motion tracking with absolute position information. In this paper, we propose a novel position estimation method which makes use of an electromagnetic resonance (EMR) motion detection board for position information to improve the tracking accuracy of a mulMU-based digital writing instrument. The EMR board cannot provide high resolution (only 3 cm per grid in our case) position information for a large writing area because of high construction cost and poor tracking performance. However, the combined scheme of using the mulMU and the EMR board can compensate their respective weaknesses. The EMR board can bound the mulMU position estimate error and the mulMU can provide detailed information of the handwriting trajectory for the rough locus obtained from the EMR board. Details of the estimation algorithm will be discussed and experimental results of its implementation are compared with the conventional Kalman filtering without the extra position feedback information.

A Novel Real-Time Error Compensation Methodology for μIMU-based Digital Writing Instrument

A micro inertial measurement unit (μlMU) which is based on Micro-Electro-Mechanical Systems (MEMS) accelerometers and gyroscope sensors is developed for real-time recognition of human hand motion. By using appropriate filtering, transformation and sensor fusion algorithms, a ubiquitous digital writing instrument is produced for recording handwriting on any surface. In this paper, we propose a method for deriving an error feedback to a Kalman filter based on the assumption that writing occurs only in two dimensions i.e. the writing surface is flat. By imposing this constraint, error feedback to the Kalman filter can be derived. Details of the feedback algorithm will be discussed and experimental results of its implementation are compared with the simple Kalman filter without feedback information.

Distinguishing cells by their first-order transient motion response under an optically induced dielectrophoretic force field

This letter reports our characterization of the transient motion of cells under an optically induced dielectrophoresis (ODEP) force field. Different types of human cells repeatably undergo a first-order transient motion response when subjected to a specific ODEP force field. A kernel function is derived to describe this transient motion. This function can be generally matched to experimental data for Raji cells and red blood cells by measuring two parameters: the initial velocity and the transient time-constant. They are uniquely different for Raji cells and RBCs. Support vector machine is used to distinguish between them based on their transient response characteristics.