G.C.T. Chow

Imperial College London, Londinium, England, United Kingdom

Are you G.C.T. Chow?

Claim your profile

Publications (19)4.27 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Proximity Query (PQ) is a process to calculate the relative placement of objects. It is a critical task for many applications such as robot motion planning, but it is often too computationally demanding for real-time applications, particularly those involving human-robot collaborative control. This paper derives a PQ formulation which can support non-convex objects represented by meshes or cloud points. We optimise the proposed PQ for reconfigurable hardware by function transformation and reduced precision, resulting in a novel data structure and memory architecture for data streaming while maintaining the accuracy of results. Run-time reconfiguration is adopted for dynamic precision optimisation. Experimental results show that our optimised PQ implementation on a reconfigurable platform with four FPGAs is 58 times faster than an optimised CPU implementation with 12 cores, 9 times faster than a GPU, and 3 times faster than a double precision implementation with four FPGAs.
    2013 International Conference on Field-Programmable Technology (FPT); 12/2013
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a real-time control framework for a snake robot with hyper-kinematic redundancy under dynamic active constraints for minimally invasive surgery. A proximity query (PQ) formulation is proposed to compute the deviation of the robot motion from predefined anatomical constraints. The proposed method is generic and can be applied to any snake robot represented as a set of control vertices. The proposed PQ formulation is implemented on a graphic processing unit, allowing for fast updates over 1 kHz. We also demonstrate that the robot joint space can be characterized into lower dimensional space for smooth articulation. A novel motion parameterization scheme in polar coordinates is proposed to describe the transition of motion, thus allowing for direct manual control of the robot using standard interface devices with limited degrees of freedom. Under the proposed framework, the correct alignment between the visual and motor axes is ensured, and haptic guidance is provided to prevent excessive force applied to the tissue by the robot body. A resistance force is further incorporated to enhance smooth pursuit movement matched to the dynamic response and actuation limit of the robot. To demonstrate the practical value of the proposed platform with enhanced ergonomic control, detailed quantitative performance evaluation was conducted on a group of subjects performing simulated intraluminal and intracavity endoscopic tasks.
    IEEE Transactions on Robotics 02/2013; 29(1):15-31. DOI:10.1109/TRO.2012.2226382 · 2.65 Impact Factor
  • Gary C. T. Chow, Wayne Luk, Philip H. W. Leong
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper introduces a novel mixed precision methodology for mathematical optimisation. It involves the use of reduced precision FPGA optimisers for searching potential regions containing the global optimum, and double precision optimisers on a general purpose processor (GPP) for verifying the results. An empirical method is proposed to determine parameters of the mixed precision methodology running on a reconfigurable accelerator consisting of FPGA and GPP. The effectiveness of our approach is evaluated using a set of optimisation benchmarks. Using our mixed precision methodology and a modern reconfigurable accelerator, we can locate the global optima 1.7 to 6 times faster compared with quad-core optimiser. The mixed precision optimisations search up to 40.3 times more starting vector per unit time compared with quad core optimisers and only 0.7% to 2.7% of these searches are refined using GPP double precision optimisers. The proposed methodology also allows us to accelerate problems with more complicated functions or to solve problems involving higher dimensions.
    Proceedings of the 2012 IEEE 20th International Symposium on Field-Programmable Custom Computing Machines; 04/2012
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: One of the main challenges when accelerating financial applications using reconfigurable hardware is the management of design complexity. This paper proposes a multi-level customisation framework for automatic generation of complex yet highly efficient curve based financial Monte Carlo simulators on reconfigurable hardware. By identifying multiple levels of functional specialisations and the optimal data format for the Monte Carlo simulation, we allow different levels of programmability in our framework to retain good performance and support multiple applications. Designs targeting a Virtex-6 SX475T FPGA generated by our framework are about 40 times faster than single-core software implementations on an i7-870 quad-core CPU at 2.93 GHz; they are over 10 times faster and 20 times more energy efficient than 4-core implementations on the same i7-870 quad-core CPU, and are over three times more energy efficient and 36% faster than a highly optimised implementation on an NVIDIA Tesla C2070 GPU at 1.15 GHz. In addition, our framework is platform independent and can be extended to support CPU and GPU applications.
    Proceedings of the 8th international conference on Reconfigurable Computing: architectures, tools and applications; 03/2012
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper introduces a novel mixed precision methodology applicable to any Monte Carlo (MC) simulation. It involves the use of data-paths with reduced precision, and the resulting errors are corrected by auxiliary sampling. An analytical model is developed for a reconfigurable accelerator system with a field-programmable gate array (FPGA) and a general purpose processor (GPP). Optimisation based on mixed integer geometric programming is employed for determining the optimal reduced precision and optimal resource allocation among the MC data-paths and correction datapaths. Experiments show that the proposed mixed precision methodology requires up to 11 % additional evaluations while less than 4 % of all the evaluations are computed in the reference precision; the resulting designs are up to 7.1 times faster and 3.1 times more energy efficient than baseline double precision FPGA designs, and up to 163 times faster and 170 times more energy efficient than quad-core software designs optimised with the Intel compiler and Math Kernel Library. Our methodology also produces designs for pricing Asian options which are 4.6 times faster and 5.5 times more energy efficient than NVIDIA Tesla C2070 GPU implementations.
    Proceedings of the ACM/SIGDA 20th International Symposium on Field Programmable Gate Arrays, FPGA 2012, Monterey, California, USA, February 22-24, 2012; 01/2012
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a generic precision optimisation methodology for quadrature computation targeting reconfigurable hardware to maximise perfor-mance at a given error tolerance level. The proposed methodology optimises per-formance by considering integration grid density versus mantissa size of floating-point operators. The optimisation provides the number of integration points and mantissa size with maximised throughput while meeting given error tolerance re-quirement. Three case studies show that the proposed reduced precision designs on a Virtex-6 SX475T FPGA are up to 6 times faster than comparable FPGA designs with double precision arithmetic. They are up to 15.1 times faster and 234.9 times more energy efficient than an i7-870 quad-core CPU, and are 1.2 times faster and 42.2 times more energy efficient than a Tesla C2070 GPU.
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Customisable data formats provide an opportunity for exploring trade-offs in accuracy and performance of reconfigurable systems. This paper introduces a novel methodology for mixed-precision comparison, which improves comparison performance by using reduced-precision data paths while maintaining accuracy by using high-precision data paths. Our methodology adopts reduced-precision data-paths for preliminary comparison, and high-precision data-paths when the accuracy for preliminary comparison is insufficient. We develop an analytical model for performance estimation of the proposed mixed-precision methodology. Optimisation based on integer linear programming is employed for determining the optimal precision and resource allocation for each of the data paths. The effectiveness of our approach is evaluated using a common collision detection problem. Performance gains of 4 to 7.3 times are obtained over baseline fixed-precision designs for the same FPGAs. With the help of the proposed mixed-precision methodology, our FPGA designs are 15.4 to 16.7 times faster than software running on multi-core CPUs with the same technology.
    Field-Programmable Custom Computing Machines (FCCM), 2011 IEEE 19th Annual International Symposium on; 06/2011
  • Source
    Adrien Le Masle, Gary Chun Tak Chow, Wayne Luk
    [Show abstract] [Hide abstract]
    ABSTRACT: We present Constant Power Reconfigurable Computing, a general and device-independent framework based on a closed-loop control system used to keep the power consumption constant for any reconfigurable computing design targeting FPGA implementation. We develop an on-chip power consumer, an on-chip power monitor and a proportional-integral-derivative controller with circuit primitives available in most commercial FPGAs. We demonstrate the effectiveness of the proposed methodology on a square-and-multiply exponentiation circuit implemented on a Spartan-6 LX45 FPGA board. By reducing the peak autocorrelation values by a factor of 2.7 on average, the proposed Constant Power Reconfigurable Computing approach decreases the information leaked by the power consumption of this system with only 26% area overhead and 28% power overhead.
    2011 International Conference on Field-Programmable Technology, FPT 2011, New Delhi, India, December 12-14, 2011; 01/2011
  • Source
    Gary Chun Tak Chow, Ken Eguro, Wayne Luk, Philip Heng Wai Leong
    [Show abstract] [Hide abstract]
    ABSTRACT: Modular multiplication of long integers is an important building block for cryptographic algorithms. Although several FPGA accelerators have been proposed for large modular multiplication, previous systems have been based on O(N2) algorithms. In this paper, we present a Montgomery multiplier that incorporates the more efficient Karatsuba algorithm which is O(N(log 3/log 2)). This system is parameterizable to different bitwidths and makes excellent use of both embedded multipliers and fine-grained logic. The design has significantly lower LUT-delay product and multiplier-delay product compared with previous designs. Initial testing on a Virtex-6 FPGA showed that it is 60-190 times faster than an optimized multi-threaded software implementation running on an Intel Xeon 2.5 GHz CPU. The proposed multiplier system is also estimated to be 95-189 times more energy efficient than the software-based implementation. This high performance and energy efficiency makes it suitable for server-side applications running in a datacenter environment.
    International Conference on Field Programmable Logic and Applications, FPL 2010, August 31 2010 - September 2, 2010, Milano, Italy; 01/2010
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A constant-power control circuit has been built successfully for the digital operation of CNT-based alcohol vapor sensors. The sensors, which are based on bundles of chemically functionalized multi-walled carbon nanotubes (f-CNTs), have been proven to be sensitive towards alcohol molecules. The resistance of the sensors increases upon exposure to alcohol vapors. The constant-power configuration is developed to avoid the self-heating effect, which is a significant factor in affecting the sensor performance. On the other hand, we also utilized the self- heating effect to clean up the alcohol molecules on the f-CNTs between measurements. The comparison experiments between constant-power and constant-current configurations were conducted. The results demonstrated larger response under constant-power mode, especially when operating power was low or alcohol concentration was relatively high. The responsivity and the sensitivity of alcohol vapor sensors under different mode and operating powers are also discussed.
    Nano/Micro Engineered and Molecular Systems, 2008. NEMS 2008. 3rd IEEE International Conference on; 02/2008
  • Gary Chun Tak Chow
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, the authors have developed an adaptive precision technique for genetic algorithms (GA) (Mitchell, 1996), which is similar in spirit to the mixed precision techniques, but instead, a global precision is used at any time. The global precision is increased adaptively, greatly simplifying the partitioning and switching decision problems.
    FPL 2008, International Conference on Field Programmable Logic and Applications, Heidelberg, Germany, 8-10 September 2008; 01/2008
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Alcohol sensors, batch fabricated by forming bundles of chemically functionalized multiwalled carbon nanotubes (f-CNTs) across Au electrodes on SiO<sub>2</sub>/Si substrates using an AC electrophoretic technique, were developed for alcohol vapor detection using an ultralow input power of ~ 0.01 - 1 muW, which is lower than the power required for most commercially available alcohol sensors by more than four orders of magnitude. The multiwalled carbon nanotubes (MWCNTs) have been chemically functionalized with the COOH groups by oxidation. We found that the sensors are selective with respect to flow from air, water vapor, and alcohol vapor. The sensor response is linear for alcohol vapor concentrations from 1 to 21 ppm with a detection limit of 0.9 ppm. The transient response of these sensors is experimentally shown to be ~1 s and the variation of the responses at each concentration is within 10% for all of the tested sensors. The sensors could also easily be reset to their initial states by annealing the f-CNTs sensing elements at a current of 100-200 muA within ~ 100-200 s. We demonstrated that the response of the sensors can be increased by one order of magnitude after adding the functional group COOH onto the nanotubes, i.e., from ~0.9% of a bare MWCNTs sensor to ~9.6% of an f-CNTs sensor with a dose of 21 ppm alcohol vapor.
    IEEE Transactions on Nanotechnology 10/2007; DOI:10.1109/TNANO.2007.900511 · 1.62 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Sensing devices using electronic-grade carbon nanotubes (EG-CNTs) as resistive sensing element were fabricated by dielectrophoresis (DEP) manipulation. DEP-based fabrication of EG-CNTs is important as it allows some control of the nominal resistance of the fabricated sensors, which is essential to improve the SNR of CNT sensors. The devices were characterized and the potential of EG-CNTs to serve as a novel temperature and humidity sensing element has been demonstrated. Electrical characterization revealed that the EG-CNTs sensors, which exhibit large linear IiquestV range, have both positive and negative TCR at higher operational temperatures. In addition, its resistance-humidity linear dependency proves its humidity sensing capability. Moreover, the EG-CNTs device is capable of operating in nW range. On the foundation of these measurements, we aim to prove EG-CNTs as a promising material for future applications in nano-sensing.
    Nanotechnology, 2007. IEEE-NANO 2007. 7th IEEE Conference on; 09/2007
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: 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.
    Robotics and Biomimetics, 2006. ROBIO '06. IEEE International Conference on; 01/2007
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We have successfully chemically functionalized the multi-walled carbon nanotubes (MWCNTs) with COOH group by the method of oxidation and used AC electrophoresis to formed these bundles MWCNTs between Au electrodes on the Si substrate. We then demonstrated that these resistive elements are capable of detecting alcohol vapor using an ultra-low input power of only ∼0.01μW. The sensors exhibit fast, repeatable, highly sensitive, and reversible response. Our results show that the resistances of the sensors vary linearly with alcohol vapor concentration from 5ppth to 100ppth (ppth = part per thousand). We can also easily reverse the initial resistance of the sensors by annealing them in real time at 100-250μA current within 1-6 minutes. We have experimental proof that the functionalized MWCNTs have a much higher sensitivity towards the alcohol vapor than the bare MWCNTs. Based on our experimental results, we prove that MWCNTs sensors, especially for those with proper functionalized groups, are sensitive to a wide range of alcohol vapor and potentially other volatile organic compounds, and are very attractive for commercialization due to their extreme low-power requirements for activation.
    Nanotechnology, 2006. IEEE-NANO 2006. Sixth IEEE Conference on; 07/2006
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper reports the characterization of a novel MEMS sensor using Carbon nanotubes (CNTs) as sensing elements. Using a MEMS-compatible process and the dielectrophoretic (DEP) nanoassembly of CNTs, we have successfully integrated bundled strands of CNT sensing elements on arrays of Polymethylmethacrylate (PMMA) diaphragms. With a thin film of Parylene-C to enhance the CNT to electrode contact, the CNT bundles show consistent and repeatable piezoresistivity from multiple electromechanical measurements. Using low noise data acquisition techniques and annealing treatments of the CNTs, the fluctuations of the sensor signal are greatly reduced. Based on these experimental evidences, we propose that embedded in Parylene-C thin film, CNT-bundles is a novel material for fabricating micro pressure sensors on polymer substrates-which may serve as ultra-low-power (μW) and low-noise sensors when bio-compatibility and low-cost applications are required.
    Nanotechnology, 2006. IEEE-NANO 2006. Sixth IEEE Conference on; 07/2006
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We have demonstrated multi-walled carbon nanotube (MWCNTs) based sensors, which are capable of detecting alcohol vapor with ultra-low power. We fabricated the Si-substrate sensors using an AC electrophoretic technique so as to form bundled MWCNTs sensing elements between Au microelectrodes. The I-V measurement illustrates that we can activate the sensors at the Ohmic region of the sensors (at 10 muA), which is without any overheat effect. The sensors only need an ultra-low power (~1 muW) to detect the alcohol vapor. They exhibit fast, reversible and repeatable response. We have tested the response of the sensors with alcohol concentrations from 10 ppth to 400 ppth (ppth = parts per thousand). Our result shows that there is a linear relation between the resistance of the sensors and alcohol concentration. Also, we can easily reverse the sensor to the initial reference resistance by annealing them at 100-250 muA current within 6 minutes. Moreover, the sensors are selective with respect to flow from air, water vapor, and alcohol vapor. Finally, we have also studied how the temperature of the sensors affects their response towards alcohol vapor. The result shows that the performance of the sensors will deteriorate as the temperature of the sensors increase. Also, the cooling effect of the vapor is not a dominating factor in determining the response of the sensor. Based on our experiments, we prove the feasibility of turning the MWCNTs sensors into a commercialized alcohol sensor with ultra-low power requirements
    Nano/Micro Engineered and Molecular Systems, 2006. NEMS '06. 1st IEEE International Conference on; 02/2006
  • [Show abstract] [Hide abstract]
    ABSTRACT: Development and preliminary characterization on the performance of a micro-bubble generator using carbon nanotubes (CNTs) as heating elements are presented. Dielectrophoretic force was used to align CNTs between fabricated micro-electrodes. The contacts between the CNTs and electrodes were fixed by patterned SiO<sub>2</sub> thin film. Localized high temperature due to Joule-heating generated by an AC current of a few hundred muA (in general les100 Hz) in the carbon nanotubes vaporized surrounding water and generated micron sized bubbles (with diameter of 5 mum detectable under normal optical microscope). These experimental evidence indicate that exposed CNTs may function as a single nucleation site for bubble generation. The power required to generate these bubbles can be as small as 337 muW, which is with 1-2 order of magnitude smaller than that of typical metal or polysilicon based MEMS heaters. Precisely controlled micro bubble generation may find promising applications in ink-jet printing, micro-fluidic visualization, or micro-fluidic based bio-manipulation
    Nano/Micro Engineered and Molecular Systems, 2006. NEMS '06. 1st IEEE International Conference on; 02/2006
  • Source
    Gary C T Chow, Ken Eguro
    [Show abstract] [Hide abstract]
    ABSTRACT: Cryptographic pairing (bilinear mapping) is a core algorithm for var-ious cryptography protocols. It is computationally expensive and inefficiently computed with general purpose processors. Although there has been previous work looking into efficient hardware designs for pairing, most of these systems use small characteristic curves which are incompatible with practical software designs. In this paper, we propose a novel processor architecture for pairing-based cryptography applications using large characteristic curves. The architec-ture is parameterizable to fields with different bit-widths and different pairing algorithms. It takes advantage of some unique FPGA features such as huge ag-gregated memory bandwidth and massively parallel computation logic to achieve high performance and high energy efficiency. An example 512-bit pair-ing processor with this architecture can verify 9.6K pairings/second on a Xilinx Virtex-6 FPGA. It is 18.7x faster than a single threaded software version run-ning on a 2.5 GHz Xeon E5420 CPU. The per-pairing energy consumption of the FPGA processor is estimated to be at least 6.0x better than its CPU counter-part. The proposed architecture is ideal for server-side applications requiring flexibility, performance and energy efficiency.