Scaling trends in microsystems are discussed frequently in the technical community, providing a short-term perspective on the future of integrated microsystems. This paper looks beyond the leading edge of technological development, focusing on new microsystem design paradigms that move far beyond today's systems based on static components. We introduce the concept of Adaptive Microsystems and outline a path to realizing these systems-on-a-chip. The role of DARPA in advancing future components and systems research is discussed, and specific DARPA efforts enabling and producing adaptive microsystems are presented. In particular, we discuss efforts underway in the DARPA Microsystems Technology Office (MTO) including programs in novel circuit architectures (3DIC), adaptive imaging and sensing (AFPA, VISA, MONTAGE, A-to-I) and reconfigurable RF/Microwave devices (SMART, TFAST, IRFFE).
It has recently been observed that sparse and compressible signals can be sketched using very few nonadaptive linear measurements in comparison with the length of the signal. This sketch can be viewed as an embedding of an entire class of compressible signals into a low-dimensional space. In particular, d-dimensional signals with
m nonzero entries (m-sparse signals) can be embedded in O(m log d) dimensions. To date, most algorithms for approximating or reconstructing the signal from the sketch, such as the linear programming approach proposed by Candes-Tao and Donoho, require time polynomial in the signal length. This paper develops a new method, called Chaining Pursuit, for sketching both m-sparse and compressible signals with O(m polylog d) nonadaptive linear measurements. The algorithm can reconstruct the original signal in time O(m polylog d) with an error proportional to the optimal m-term approximation error. In particular, m-sparse signals are recovered perfectly and compressible signals are recovered with polylogarithmic distortion. Moreover, the algorithm can operate in small space O(m polylog d), so it is appropriate for streaming data.
Compressive sampling (CS), or Compressed Sensing, has generated a tremendous amount of excitement in the signal processing community. Compressive sampling, which involves non-traditional samples in the form of randomized projections, can capture most of the salient information in a signal with a relatively small number of samples, often far fewer samples than required using traditional sampling schemes. Adaptive sampling (AS), also called Active Learning, uses information gleaned from previous observations (e.g., feedback) to focus the sampling process. Theoretical and experimental results have shown that adaptive sampling can dramatically outperform conventional (non-adaptive) sampling schemes. This paper compares the theoretical performance of compressive and adaptive sampling for regression in noisy conditions, and it is shown that for certain classes of piecewise constant signals and high SNR regimes both CS and AS are near optimal. This result is remarkable since it is the first evidence that shows that compressive sampling, which is non-adaptive, cannot be significantly outperformed by any other method (including adaptive sampling procedures), even in the presence of noise. The performance of CS schemes for signal detection is also investigated.
We describe signal processing tools to extract structure and information from arbitrary digital data sets. In particular heterogeneous multi-sensor measurements which involve corrupt data, either noisy or with missing entries present formidable challenges. We sketch methodologies for using the network of inferences and similarities between the data points to create robust nonlinear estimators for missing or noisy entries. These methods enable coherent fusion of data from a multiplicity of sources, generalizing signal processing to a non linear setting. Since they provide empirical data models they could also potentially extend analog to digital conversion schemes like "sigma delta".
This paper describes a compressive sensing strategy developed under the Compressive Optical MONTAGE Photography Initiative. Multiplex and multi-channel measurements are generally necessary for compressive sensing. In a compressive imaging system described here, static focal plane coding is used with multiple image apertures for non-degenerate multiplexing and multiple channel sampling. According to classical analysis, one might expect the number of pixels in a reconstructed image to equal the total number of pixels across the sampling channels, but we demonstrate that the system can achieve up to 50% compression with conventional benchmarking images. In general, the compression rate depends on the compression potential of an image with respect to the coding and decoding schemes employed in the system.
We consider compressive sensing in the context of optical spectroscopy. With compressive sensing, the ratio between the number of measurements and the number of estimated values is less than one, without compromising the fidelity in estimation. A compressive sensing system is composed of a measurement subsystem that maps a
signal to digital data and an inference algorithm that maps the data to a signal estimate. The inference algorithm exploits both the information captured in the measurement and certain a priori information about the signals of interest, while the measurement subsystem provides complementary, signal-specific information at the lowest sampling rate possible. Codesign of the measurement strategies, the model of a priori information, and the
inference algorithm is the central problem of system design. This paper describes measurement constraints specific to optical spectrometers, inference models based on physical or statistical characteristics of the signals, as well as linear and nonlinear reconstruction algorithms. We compare the fidelity of sampling and inference strategies over a family of spectral signals.
Creating intelligent integrated microsystems, devices that incorporate photonics, electronics, MEMS, and embedded intelligence, presents multiple challenges. The three device technologies have been largely developed independently and have established their own sets of design and process rules that have led to highly stable, high yield processes. In combining these technologies to achieve a desired functionality, constraints are placed on each technology to avoid adverse impacts on the others. Finding a common path towards achieving a single end objective requires process reoptimization and the development of new processes. This paper discusses the dual band adaptive focal plane array (AFPA) that is currently under development, with an emphasis on technology integration and the resulting functional benefits that can be realized. The AFPA device is a dual band IR imaging sensor that enables simultaneous collection of high-resolution MWIR imagery, with spatially independent spectrally tuned imaging in the LWIR for enhanced target detection and classification.
Hyperspectral imaging in the infrared bands is traditionally performed using a broad spectral response focal plane array, integrated in a grating or a Fourier transform spectrometer. This paper describes an approach for miniaturizing a hyperspectral detection system on a chip by integrating a Micro-Electro-Mechanical-System (MEMS) based tunable Fabry Perot (FP) filter directly on a photodetector. A readout integrated circuit (ROIC) serves to both integrate the detector signal as well as to electrically tune the filter across the wavelength band. We report the first such demonstration of a tunable MEMS filter monolithically integrated on a HgCdTe detector. The filter structures, designed for operation in the 1.6-2.5 μm wavelength band, were fabricated directly on HgCdTe detectors, both in photoconducting and high density vertically integrated photodiode (HDVIP) detectors. The HDVIP detectors have an architecture that permits operation in the standard photodiode mode at low bias voltages (≤0.5V) or in the electron avalanche photodiode (EAPD) mode with gain at bias voltages of ~20V. In the APD mode gain values of 100 may be achieved at 20 V at 200 K. The FP filter consists of distributed Bragg mirrors formed of Ge-SiO-Ge, a sacrificial spacer layer within the cavity and a silicon nitride spacer membrane for support. Mirror stacks fabricated on silicon, identical to the structures that will form the optical cavity, have been characterized to determine the optimum filter characteristics. The measured full width at half maximum (FWHM) was 34 nm at the center wavelength of 1780 nm with an extinction ratio of 36.6. Fully integrated filters on HgCdTe photoconductors with a center wavelength of approximately 1950 nm give a FWHM of approximately 100 nm, and a peak responsivity of approximately 8×104 V/W. Initial results for the filters on HDVIP detectors exhibit FWHM of 140 nm.
The ever increasing demand for more electronic functionality integrated in silicon chips requires new approaches. One solution is to explore several technology fronts such as nanometer lithography or advanced materials. A better (but complementary) solution is to use the third dimension by stacking multiple layers of integrated circuits that overcomes the limitations of planar approaches. A general overview of 3D packaging and integration approaches is presented in relation to system architectures that will benefit from 3D implementation.
We propose the design and fabrication of nanophotonic optical routing channels using three-dimensional photonic crystals (PhCs) operating at telecommunication wavelengths. The fabrication method involves patterning a single planar etch mask and custom-tuned isotropic reactive-ion etching with passivation to create an array of spherical voids with three-dimensional symmetry. Introducing planar defects in the buried PhC lattice can create the optical routing channels. The dispersion properties of the embedded silicon three-dimensional (3D) PhC are utilized for the optical interconnect
design. The proposed design has sub-micron routing capability and flexibility with overlay and arbitrary routing. The optoelectronic circuits can be fabricated using CMOS technology on the surface and the source and emitters can be flip chip bonded to emit down, through the silicon layer. The beam coupling can be done using angled mirror facets. The guiding mechanism is based on the three-dimensional self-collimation effect.
Within the government communication market there is an increasing push to further miniaturize systems with the use of chip-scale packages, flip-chip bonding, and other advances over traditional packaging techniques. Harris' approach to miniaturization includes these traditional packaging advances, but goes beyond this level of miniaturization by combining the functional and structural elements of a system, thus creating a Multi-Functional Structural Circuit
(MFSC). An emerging high-frequency, near hermetic, thermoplastic electronic substrate material, Liquid Crystal Polymer (LCP), is the material that will enable the combination of the electronic circuit and the physical structure of the system. The first embodiment of this vision for Harris is the development of a battlefield acoustic sensor module. This paper will introduce LCP and its advantages for MFSC, present an example of the work that Harris has performed, and
speak to LCP MFSCs' potential benefits to miniature communications modules and sensor platforms.
Conventional imaging systems can suffer from significant aliasing and/or blur distortions when the detector array in the focal plane under-samples the image. We propose to address this problem by engineering the optical PSF of the imaging system followed by electronic post-processing to minimize the overall distortions. The optical PSF of the candidate imaging system is modified by placing a phase-mask in its aperture-stop. We consider a particular parameterization of the phase-mask and optimize its parameters to minimize the distortions. We obtain as much as 30% improvement in the final imaging quality with the optimized optical PSF imager (SPEL) relative to the conventional imager.
We analyze the performance of a novel detector array for detecting and localizing particle emitting sources. The array is spherically shaped and consists of multiple "eyelets," each having a conical shape with a lens on top and a particle detectors subarray inside. The array's configuration is inspired by and generalizes the biological compound eye: it has a global spherical shape and allows a large number of detectors in each eyelet. The array can be used to detect particles including photons (e.g. visible light, X or γ rays), electrons, protons, neutrons, or α particles. We analyze the performance of the array by computing statistical Cramer-Rao bounds on the errors in estimating the direction of arrival (DOA) of the incident particles. In numerical examples, we first show the influence of the array parameters on its performance bound on the mean-square angular error (MSAE). Then we optimize the array's configuration according to a min-max criterion, i.e. minimize the worst case lower bound of the MSAE. Finally we introduce two estimators of the source direction using the proposed array and analyze their performance, thereby showing that the performance bound is attainable in practice. Potential applications include artificial vision, astronomy, and security.
We present room-temperature AlGaAsSb/InGaAsSb heterojunction phototransistors (HPT) with a cutoff wavelength (50% of maximum quantum efficiency) of 2.4 μm and 2.15 μm. AlGaAsSb/InGaAsSb HPT structures were grown by molecular beam epitaxy (MBE) or metal-organic chemical vapor deposition (MOCVD). This work is a continuation of a preceding project, which was carried out using liquid phase epitaxy (LPE) grown AlGaAsSb/InGaAsSb/GaSb heterostructures. Although the LPE-related work resulted in the fabrication of an HPT with excellent parameters, MBE and MOCVD - compared to LPE - provides better control over doping levels, composition and width of the AlGaAsSb and InGaAsSb layers, compositional and doping profiles, especially with regard to abrupt heterojunctions. HPT with different diameter of photosensitive area (75, 200, 300 and 1000 μm) were fabricated and characterized. In particular, I-V characteristics, spectral response and noise, as well as detectivity and noise-equivalent-power were determined in a broad range of temperatures and bias voltages. Advantages of HPT integration with diffractive optical elements (DOE) were demonstrated.
In this paper we present several novel photonic technologies for sensing millimeter-wave (MMW) radiation for the imaging and spectroscopy applications. Based on the optical up-conversion approach, our high-sensitivity MMW imaging system transfers the power of MMW radiation received from a broadband horn antenna to the sidebands on an optical carrier via an electrooptic (EO) modulator. The detection is realized by measuring the transferred optical power of the sidebands. The sensitivity of this detection system is primarily controlled by the conversion efficiency of the EO modulator
at the desired MMW frequency. In this paper, we present the design, fabrication, and characteristics of the ultra-broadband LiNbO3 traveling-wave modulator for the MMW detection system working at a frequency of 95 GHz. A numerical model based on the finite element analysis technique has developed to optimize the device geometric parameters and the fabrication processes. A modulation efficiency of ~0.9 W-1 at 95 GHz has been achieved for the optimized modulator, which corresponds to the half-wave voltages of 9 V and 18 V, at DC and 95 GHz, respectively. The detection pixel based on those modulators has shown a high sensibility with a noise equivalent temperature difference of ~17K at a refreshing rate of 30 Hz.
We present the design and implementation of an ultra-thin camera using annular folded optics. An 8-fold optic is fabricated from a single piece of Calcium Fluoride; and demonstrated experimentally to have a 35 mm effective focal length, 0.12 radian field of view, 0.07 mrad resolution, 27 mm effective aperture diameter and 5 mm total thickness.
Architectures are discussed in which power amplifiers are able to vary their performance characteristics in response to changing output power levels, frequencies, load impedance levels and linearity constraints. Necessary elements for implementation of such systems are reviewed, including sensors to detect changes in the external environment, actuators to adapt the amplifier characteristics, and algorithms to implement the associated control functions. An example is described of a power amplifier for cell phone applications that measures the load impedance provided by the antenna and varies its output impedance match accordingly, in order to preserve output power, efficiency and linearity as the external antenna is handled. The system automatically provides output tuning for a 1W linear amplifier, and can accommodate output standing wave ratios up to 8:1. It improves output power and efficiency by x2 in representative mismatch scenarios. The insertion loss of the system is comfortably low at 0.5dB.
We present the potential of using Programmable Analog Signal processing techniques for impacting low-power portable applications like imaging, audio processing, and speech recognition. The range of analog signal processing functions available results in many potential opportunities to incorporate these analog signal processing systems with digital signal processing systems for improved overall system performance. We describes our programmable analog technology based around floating-gate transistors that allow for non-volitile storage as well as computation through the same device. We describe the basic concepts for floating-gate devices, capacitor-based circuits, and the basic charge modification mechanisms that makes this analog technology programmable. We describes the techniques to extend these techniques to program an array of floating-gate devices. We show experimental evidence for the factor of 1000 to 10,000 power efficiency improvement for programmable analog signal processing compared to custom digital implementations in Vector Matrix Multipliers (VMM), CMOS imagers with computation on the pixel plane with high fill factors, and Large-Scale Field Programmable Analog Arrays (FPAA), among others.
Progress-to-date of a S-Ku band intelligent amplifier microsystem is presented. Performance objectives are 0.5 Watt with 30%-55% power added efficiency across the band using a total of 10 RF MEMS. GaAs-to-GaAs and Borosilicate-to-GaAs low temperature (<250C) indium-gold wafer bonding is employed to provide hermeticity and integration of the pHEMT transistor with the RF MEMS. The compact mixed signal microsystem, 2 inches by 3 inches, utilizes an existing data processor with an intelligent control algorithm which optimizes the input and output circuit matching networks for optimal performance.