Property:Abstract
From Murray Wiki
Jump to navigationJump to search
This is a property of type Text.
A
The bootstrapping problem consists in designing agents that laern a model of themsleves and the world, and utilize it to achieve useful tasks. It is different from other learning problems as the agent starts with uninterpreted observaions and commands, and with minimal prior information about the world. In this paper, we give a mathematical formalizatoin of this aspect of the problem. We argue that the vague constraint of having âno prior informationâ can be recast as a precise algebraic condition on the agent: that its behavior is invariant to particular classes of nuisances on the world, which we show can be well represented by actions of groups (diffeomorphisms, permutatations, linear transformations) on observations and commands. We then introduce the class of bilinear gradient dynamics sensors (DGDS) as a candidate for learning generic rootic sensorimotor cascades. We show how framing the problem as rejections of group nuisances allows a compact and modular analysis of typical preprocessing stages, such as learning the toplogy of sensors. We demonstrate learning and using such models on real-world range-finder and camera data from publicly available datasets. +
C
The bottom up design of genetic circuits to control cellular behavior is one of the central objectives within Synthetic Biology. Performing design iterations on these circuits in vivo is often a time consuming process, which has led to E. coli cell extracts to be used as simplified circuit prototyping environments. Cell extracts, however, display large batch-to-batch variability in gene expression. In this paper, we develop the theoretical groundwork for a model based calibration methodology for correcting this variability. We also look at the interaction of this methodology with the phenomenon of parameter (structural) non-identifiability, which occurs when the parameter identification inverse problem has multiple solutions. In particular, we show that under certain consistency conditions on the sets of output- indistinguishable parameters, data variability reduction can still be performed, and when the parameter sets have a cer- tain structural feature called covariation, our methodology may be modified in a particular way to still achieve the desired variability reduction. +
P
The control programs of complex autonomous systems that have conditional branching can be modeled as linear hybrid systems. When the state knowledge is perfect, linear hybrid systems with state-based transition conditions can be verified against a specified unsafe set using existing model checking software. This paper introduces a formal method for calculating the failure probability due to state estimation uncertainty of these sensor-driven hybrid systems. Problem complexity is described and some reduction techniques for the failure probability calculation are given. An example goal-based control program is given and the failure probability for that system is calculated. +
J
The cost of the great expressivity of motion planning subject to temporal logic formulae is intractability. Recent advances in sampling-based methods seem to be only applicable to âlow-levelâ control. The problem of realizing âhigh-levelâ controllers that satisfy a temporal logic specification does not readily admit approximations, unless the notion of correctness is relaxed as might be achieved with probabilistic variants of temporal logics. In this paper, we argue that not all possible environment (uncontrolled) behaviors need to be explicitly planned for, but rather short-time strategies can be generated online while maintaining global correctness. We achieve this by separating feasibility from controller synthesis, using a metric from the underlying continuous state space to ensure short-time strategies chained together provide globally correct behavior. +
A
The development of autonomous vehicles for urban driving has seen rapid progress in the past 30 years. This paper provides a summary of the current state of the art in autonomous driving in urban environments, based primarily on the experiences of the authors in the 2007 DARPA Urban Challenge (DUC). The paper briefly summarizes the approaches that different teams used in the DUC, with the goal of describing some of the challenges that the teams faced in driving in urban environments. The paper also highlights the long term research challenges that must be overcome in order to enable autonomous driving and points to opportunities for new technologies to be applied in improving vehicle safety, exploiting intelligent road infrastructure and enabling robotic vehicles operating in human environments. +
O
The fact that genes compete for shared cellular resources poses a fundamental challenge when identifying pa- rameters of synthetic gene circuits. A recently developed model of gene expression tackles this problem by explicitly accounting for resource competition. In addition to accurately describing experimental data, this model only depends on a handful of easily identifiable parameters with clear physical interpretation. Based on this model, we outline a procedure how to select the optimal set of experiments to characterize biomolecular parts in synthetic biology. Additionally, we reveal the role competition for shared resources plays, provide guidelines how to minimize its detrimental effects, and how to leverage this phenomenon to extract the most information about unknown parameters. To illustrate the results, we consider the case of part characterization in cell-free extracts, treat plasmid DNA concentrations as decision variables, and demonstrate the significant performance difference between na ̈ıve and optimal experiment design. +
F
The field of control provides the principles and methods used to design
physical and information systems that maintain desirable performance by sensing
and automatically adapting to changes in the environment. As we begin the 21st
Century, the opportunities to apply control principles and methods are exploding.
New developments in this increasingly information rich world will require a significant
expansion of the basic tool sets of control. This article summarizes the main
findings and recommendations of the Panel on Future Directions in Control, Dynamics,
and Systems, which has recently released its report. The report spells out some
of the prospects for control in the current and future technological environment,
describes the role the field will play in military, commercial, and scientific
applications over the next decade, and recommends actions required to enable new
breakthroughs in engineering and technology through application of control research. +
M
The focus of this paper is on modeling and control of Smart Thermal Grids (STGs) in which the uncertainties in the demand and/or supply are included. We solve the corre- sponding robust model predictive control (MPC) optimization problem using mixed-integer-linear programming techniques to provide a day-ahead prediction for the heat production in the grid. In an example, we compare the robust MPC approach with the robust optimal control approach, in which the day-ahead production plan is obtained by optimizing the objective function for entire day at once. There, we show that the robust MPC approach successfully keeps the supply-demand balance in the STG while satisfying the constraints of the production units in the presence of uncertainties in the heat demand. Moreover, we see that despite the longer computation time, the performance of the robust MPC controller is considerably better than the one of the robust optimal controller. +
P
The goal of this work is to provide accurate dynamical models of oscillations in the flow past a rectangular cavity, for the purpose of bifurcation analysis and control. We have performed an extensive set of direct numerical simulations which provide the data used to derive and evaluate the models. Based on the method of Proper Orthogonal Decomposition (POD) and Galerkin projection, we obtain low-order models (from 6 to 60 states) which capture the dynamics very accurately over a few periods of oscillations, but deviate for long time. +
R
The increasing complexity of electric power sys- tems leads to integration and verification challenges. We consider the problem of designing a control protocol for the aircraft electric power system that meets these system requirements and reacts dynamically to changes in internal system states. We formalize these requirements by translating them into a temporal logic specification language describing the correct behaviors of the system, and apply formal methods to automatically synthesize a controller protocol that satisfies these overall properties and requirements. Through an example, we perform a design exploration to show the benefits and tradeoffs between centralized and distributed control architectures. +
C
Converting Smooth, Time-Varying, Asymptotic Stabilizers for Driftless Systems to Homogeneous, Exponential Stabilizers +
The main result of this paper is a theorem that allows smooth,
time-varying controllers which asymptotically stabilize a driftless
nonlinear system to be converted to homogeneous, time-varying
controllers which provide {\em exponential} stability. The resulting
controllers are smooth everywhere except the origin and are easily
computed given the original asymptotic stabilizer. We illustrate the
result with experimental results on a simple mobile robot. +
I
The malachite green aptamer (MGapt) is known for its utility in RNA measurement in vivo and lysate-based cell-free protein systems. However, MGapt fluorescence dynamics do not accurately reflect mRNA concentration. Our study finds that MGapt fluorescence is unstable in commercial PURE systems. We discovered that the chemical composition of the cell-free reaction strongly influences MGapt fluorescence, which leads to inaccurate RNA calculations. Specific to the commercial system, we posit that MGapt fluorescence is significantly affected by the system’s chemical properties, governed notably by the presence of dithiothreitol (DTT). We propose a model that, on average, accurately predicts MGapt measurement within a 10% margin, leveraging DTT concentration as a critical factor. This model sheds light on the complex dynamics of MGapt in cell-free systems and underscores the importance of considering environmental factors in RNA measurements using aptamers. +
E
The mammalian gut contains trillions of microbes that interact with host cells and monitor changes in the environment. Opportunistic pathogens exploit environmental conditions to stimulate their growth and virulence, leading to a resurgence of chronic disorders such as inflammatory bowel disease (IBD). Current therapies are effective in less than 30% of patients due to the lack of adherence to prescription schedules and overall, off-target effects. Smart microbial therapeutics can be engineered to colonize the gut, providing in situ surveillance and conditional disease modulation. However, many current engineered microbes can only respond to single gut environmental factors, limiting their effectiveness. In this work, we implement the previously characterized split activator AND logic gate in the probiotic E. coli strain Nissle 1917. Our system can respond to two input signals: the inflammatory biomarker tetrathionate and a second input signal, IPTG. We report 4-6 fold induction with minimal leak when both signals are present. We model the dynamics of the AND gate using chemical reaction networks, and by tuning parameters in silico, we identified perturbations that affect our circuit's selectivity. We anticipate that our results will prove useful for designing living therapeutics for spatial targeting and signal processing in complex environments. +
R
The many successes of synthetic biology have come in a manner largely different from those in other en- gineering disciplines; in particular, without well-characterized and simplified prototyping environments to play a role analogous to wind-tunnels in aerodynamics and breadboards in electrical engineering. However, as the complexity of synthetic circuits increases, the benefitsÂin cost savings and design cycle time of a more traditional engineering approach can be significant. We have recently developed an in vitro Âbreadboard prototyping platform based on E. coli cell extract that allows biocircuits to operate in an environment considerably simpler than but functionally similar to in vivo. The simplicity of the cell-free transcription-translation breadboard makes it a promising tool for rapid biocircuit design and testing, as well as for probing the fundamentals of gene circuit functions that are normally masked by cellular complex- ity. In this work we characterize the cell-free breadboard using real-time and simultaneous measurements of transcriptional and translational activities of a small set of reporter genes and a transcriptional activation cascade. We determine the effects of promoter strength, gene and nucleoside triphosphate concentrations on biocircuits properties, and we isolate contributions of the essential componentsÂcore RNA polymerase, housekeeping sigma factor, and ribosomesÂto overall performance. Importantly, we show how limits on essential resources, particularly those involved in translation steps, manifest themselves in the form of reduced expression in the presence of orthogonal genes as load processes. +
E
Experimental Demonstration of the Dynamics and Stability of a Low Reynolds Number Swimmer Near a Plane Wall +
The motion of microorganisms as well as of tiny robotic swimmers for biomedical applications is governed by low Reynolds number (Re) hydrodynamics, where viscous effects dominate and inertial effects are negligible. This paper presents experimental results that verify theoretical predictions of our recent work which analyzed the dynamics and stability of a low-Re swimmer near a plane wall. The experimental setup uses macro-scale swimmer prototypes which are propelled by rotating cylinders in highly viscous silicone oil. The motion was recorded by a video camera and position measurements were taken by an optical tracking system. The results show good qualitative agreement with our recent theoretical predictions. +
S
The objective of this work is to study the benefits that delay can provide in simplifying the control process of large-scale systems, motivated by the availability of different types of delays in man-made and biological systems. We show that a continuous-time linear time-invariant (LTI) controller can be approximated by a simple controller that mainly uses delay blocks instead of integrators. More specifically, three methods are proposed to approximate a pre-designed stabilizing LTI controller arbitrarily precisely by a simple delay-based controller composed of delay blocks, a few integrators and possibly a unity feedback. Different problems associated with the developed approximation procedures, such as finding the optimal number of delay blocks or studying the robustness of the designed controller with respect to delay values, are then addressed. +
O
The observability properties of a class of hybrid
systems whose continuous variables are available for measurement
are considered. We show that the discrete variables'
dynamics can be always extended for observable systems to
a lattice in such a way that the extended system has the
properties that allow the construction of the LU discrete
state estimator. Such an estimator updates two variables at
each step, namely the upper and lower bound of the set
of all possible discrete variables' values compatible with the
output sequence. We give an estimate of the complexity of the
estimator in the worst case. +
I
Implementation And Simulation Of Phosphorylation-Based Insulator In Transcription-Translation Platform +
The operational amplifier (OPAMP) is a very useful insulation module in electric circuits to avoid loading effect (retroactivity). In synthetic biological circuits, we also have the same retroactivity problem, in which the biomolecular systems are not always modular due to downstream components. The output of the upstream component will be affected as the downstream component sequesters that output, which in turn impedes the process of constructing more complex biocircuits. To address this obstacle, the retroactivity needs to be attenuated by implementing a similar OPAMP device using biocircuits. Previous theoretical papers suggested a potential function of a phosphorylation based circuit in providing the feature of atten- uating retroactivity. Here we presented a successful prototyping and implementation of such a phosphorylation-based insulator (PBI) in an in vitro cell-free transcription-translation system (TX- TL). We demonstrated that retroactivity also exists in TX-TL system, if not stronger, by testing a simple negative regulation circuit. Besides we showed that the TX-TL system contains all the protein, DNA components and other resources required for the PBI circuit to work properly. We then demonstrated that the PBI circuit helps minimizing the loading effect to less than 10% compared to control circuit. With this preliminary PBI circuit design, attenuation of retroactivity while connecting two modules in vitro becomes possible. In concert with another paper from our group (E. Yeung, S. Guo, R. Murray QBIO2014) which used system identification to estimate all three essential parameters in a simplified PBI model, we showed that the simulations based on these parameters match the experimental data very well and provide an interesting insight into future designs. +
G
The paper studies the problem which we refer to as geometric trajectory filtering, where only trajectories that satisfy the local safety constraints are selected from a library of trajectories. The goal is to speed up primitive-based motion planning while still maintaining a relatively a large collection of motion primitives. One way to solve this problem is to obtain a proper (preferably smooth) function, referred to as the containment indicator function, that describes the shape of the free space. To construct the containment indicator function for an arbitrary shape, the paper uses conformal mapping to transform the original shape of interest to a simpler target shape (e.g. disk, rectangle), which can then be characterized by elementary functions. Computational methods for finding the desired conformal maps are studied. It is shown that they can be formulated as convex optimization problems, whose solution can be obtained efficiently. +
L
The problem of bootstrapping consists in designing agents that can learn from scratch the model of their sensori- motor cascade (the series of robot actuators, the external world, and the robot sensors) and use it to achieve useful tasks. In principle, we would want to design agents that can work for any robot dynamics and any robot sensor(s). One of the difficulties of this problem is the fact that the observations are very high dimensional, the dynamics is nonlinear, and there is a wide range of "representation nuisances" to which we would want the agent to be robust. In this paper, we model the dynamics of sensorimotor cascades using diffeomorphisms of the sensel space. We show that this model captures the dynamics of camera and range-finder data, that it can be used for long-term predictions, and that it can capture nonlinear phenomena such as a limited field of view. Moreover, by analyzing the learned diffeomorphisms it is possible to recover the "linear structure" of the dynamics in a manner which is independent of the commands representation. +
