Andrew Tzer-Yeu Chen

Use of Temperature and Humidity Sensors to Determine Moisture Content of Oolong Tea

The measurement of tea moisture content is important for processing and storing tea. The moisture content of tea affects the quality and durability of the product. Some electrical devices have been proposed to measure the moisture content of tea leaves but are not practical. Their performance is influenced by material density and packing. The official oven method is time-consuming. In this study, the moisture content of Oolong tea was measured by the equilibrium relative humidity technique. The equilibrium relative humidity, and temperature, of tea materials were measured by using temperature and relative humidity sensors. Sensors were calibrated, and calibration equations were established to improve accuracy. The moisture content was calculated by using an equilibrium moisture content model. The error of the moisture content determined with this method was within 0.5% w.b. at moisture <15% w.b. Uncertainty analysis revealed that the performance of the humidity sensor had a significant effect on the accuracy of moisture determination.

Evaluation of Piecewise Polynomial Equations for Two Types of Thermocouples

Thermocouples are the most frequently used sensors for temperature measurement because of their wide applicability, long-term stability and high reliability. However, one of the major utilization problems is the linearization of the transfer relation between temperature and output voltage of thermocouples. The linear calibration equation and its modules could be improved by using regression analysis to help solve this problem. In this study, two types of thermocouple and five temperature ranges were selected to evaluate the fitting agreement of different-order polynomial equations. Two quantitative criteria, the average of the absolute error values |e|ave and the standard deviation of calibration equation estd, were used to evaluate the accuracy and precision of these calibrations equations. The optimal order of polynomial equations differed with the temperature range. The accuracy and precision of the calibration equation could be improved significantly with an adequate higher degree polynomial equation. The technique could be applied with hardware modules to serve as an intelligent sensor for temperature measurement.

Convolutional neural network acceleration with hardware/software co-design

Convolutional Neural Networks (CNNs) have a broad range of applications, such as image processing and natural language processing. Inspired by the mammalian visual cortex, CNNs have been shown to achieve impressive results on a number of computer vision challenges, but often with large amounts of processing power and no timing restrictions. This paper presents a design methodology for accelerating CNNs using Hardware/Software Co-design techniques, in order to balance performance and flexibility, particularly for resource-constrained systems. The methodology is applied to a gender recognition case study, using an ARM processor and FPGA fabric to create an embedded system that can process facial images in real-time.

Computer vision based chess playing capabilities for the Baxter humanoid robot

This paper presents a project that allows the Baxter humanoid robot to play chess against human players autonomously. The complete solution uses three main subsystems: computer vision based on a single camera embedded in Baxters arm to perceive the game state, an open-source chess engine to compute the next move, and a mechatronics subsystem with a 7-DOF arm to manipulate the pieces. Baxter can play chess successfully in unconstrained environments by dynamically responding to changes in the environment. This implementation demonstrates Baxters capabilities of vision-based adaptive control and small-scale manipulation, which can be applicable to numerous applications, while also contributing to the computer vision chess analysis literature.

A Single RF Emitter-Based Indoor Navigation Method for Autonomous Service Robots

Location-aware services are one of the key elements of modern intelligent applications. Numerous real-world applications such as factory automation, indoor delivery, and even search and rescue scenarios require autonomous robots to have the ability to navigate in an unknown environment and reach mobile targets with minimal or no prior infrastructure deployment. This research investigates and proposes a novel approach of dynamic target localisation using a single RF emitter, which will be used as the basis of allowing autonomous robots to navigate towards and reach a target. Through the use of multiple directional antennae, Received Signal Strength (RSS) is compared to determine the most probable direction of the targeted emitter, which is combined with the distance estimates to improve the localisation performance. The accuracy of the position estimate is further improved using a particle filter to mitigate the fluctuating nature of real-time RSS data. Based on the direction information, a motion control algorithm is proposed, using Simultaneous Localisation and Mapping (SLAM) and A* path planning to enable navigation through unknown complex environments. A number of navigation scenarios were developed in the context of factory automation applications to demonstrate and evaluate the functionality and performance of the proposed system.

Trusting the Computer in Computer Vision: A Privacy-Affirming Framework

The use of surveillance cameras continues to increase, ranging from conventional applications such as law enforcement to newer scenarios with looser requirements such as gathering business intelligence. Humans still play an integral part in using and interpreting the footage from these systems, but are also a significant factor in causing unintentional privacy breaches. As computer vision methods continue to improve, we argue in this position paper that system designers should reconsider the role of machines in surveillance, and how automation can be used to help protect privacy. We explore this by discussing the impact of the human-in-the-loop, the potential for using abstraction and distributed computing to further privacy goals, and an approach for determining when video footage should be hidden from human users. We propose that in an ideal surveillance scenario, a privacy-affirming framework causes collected camera footage to be processed by computers directly, and never shown to humans. This implicitly requires humans to establish trust, to believe that computer vision systems can generate sufficiently accurate results without human supervision, so that if information about people must be gathered, unintentional data collection is mitigated as much as possible.

Accelerating SuperBE with Hardware/Software Co-Design

Background Estimation is a common computer vision task, used for segmenting moving objects in video streams. This can be useful as a pre-processing step, isolating regions of interest for more complicated algorithms performing detection, recognition, and identification tasks, in order to reduce overall computation time. This is especially important in the context of embedded systems like smart cameras, which may need to process images with constrained computational resources. This work focuses on accelerating SuperBE, a superpixel-based background estimation algorithm that was designed for simplicity and reducing computational complexity while maintaining state-of-the-art levels of accuracy. We explore both software and hardware acceleration opportunities, converting the original algorithm into a greyscale, integer-only version, and using Hardware/Software Co-design to develop hardware acceleration components on FPGA fabric that assist a software processor. We achieved a 4.4× speed improvement with the software optimisations alone, and a 2× speed improvement with the hardware optimisations alone. When combined, these led to a 9× speed improvement on a Cyclone V System-on-Chip, delivering almost 38 fps on 320 × 240 resolution images.

A framework for designing dynamic and interoperable automation and robotics systems

In this paper, we propose using an approach based on the system-level programming language SystemJ extended with service oriented features, called SOSJ, to design dynamic interoperable software systems. The approach abstracts and integrates the worlds of automation and robotics systems by using a simple service interface based on abstract objects within SOSJ. We demonstrate our approach in a real-life automated bottling system scenario that uses multiple FESTO modular stations operating in SOSJ and integrating them with two Baxter robots operating in ROS without the need for any modification of the underlying mechatronics or robotics systems.

Designing Dynamic and Collaborative Automation and Robotics Software Systems

The heterogeneity of execution platforms and operating software in manufacturing machines and robots, as well as various sensors and actuators, creates challenges for integration into larger systems. Existing approaches make use of different types of middleware to mitigate the challenges of designing interoperable systems. However, middleware can significantly impede modular design and composition of software systems that are dynamic in nature. This paper elaborates upon those challenges and proposes using an approach called service-oriented SystemJ (SOSJ), based on the system-level programming language SystemJ enhanced with service oriented features. This approach allows developers to design dynamic software systems while adopting and incorporating legacy solutions. The approach is demonstrated on the integration of an industrial automation system, incorporating the use of multiple modular mechatronics stations and service robotics systems, represented by robot operating system-enabled Baxter robots. The proposed approach offers a simple service interface based on abstract objects for integrating robots and automation machines in the SOSJ world, without the need to modify the underlying mechatronics or robotics systems.

Hardware/Software Co-design for a Gender Recognition Embedded System

Gender recognition has applications in human-computer interaction, biometric authentication, and targeted marketing. This paper presents an implementation of an algorithm for binary male/female gender recognition from face images based on a shunting inhibitory convolutional neural network, which has a reported accuracy on the FERET database of 97.2 %. The proposed hardware/software co-design approach using an ARM processor and FPGA can be used as an embedded system for a targeted marketing application to allow real-time processing. A threefold speedup is achieved in the presented approach compared to a software implementation on the ARM processor alone.