Theodore P. Pavlic

Design of ant-inspired stochastic control policies for collective transport by robotic swarms

In this paper, we present an approach to designing decentralized robot control policies that mimic certain microscopic and macroscopic behaviors of ants performing collective transport tasks. In prior work, we used a stochastic hybrid system model to characterize the observed team dynamics of ant group retrieval of a rigid load. We have also used macroscopic population dynamic models to design enzyme-inspired stochastic control policies that allocate a robotic swarm around multiple boundaries in a way that is robust to environmental variations. Here, we build on this prior work to synthesize stochastic robot attachment–detachment policies for tasks in which a robotic swarm must achieve non-uniform spatial distributions around multiple loads and transport them at a constant velocity. Three methods are presented for designing robot control policies that replicate the steady-state distributions, transient dynamics, and fluxes between states that we have observed in ant populations during group retrieval. The equilibrium population matching method can be used to achieve a desired transport team composition as quickly as possible; the transient matching method can control the transient population dynamics of the team while driving it to the desired composition; and the rate matching method regulates the rates at which robots join and leave a load during transport. We validate our model predictions in an agent-based simulation, verify that each controller design method produces successful transport of a load at a regulated velocity, and compare the advantages and disadvantages of each method.

Foraging theory for autonomous vehicle speed choice

We consider the optimal control design of an abstract autonomous vehicle (AAV). The AAV searches an area for tasks that are detected with a probability that depends on vehicle speed, and each detected task can be processed or ignored. Both searching and processing are costly, but processing also returns rewards that quantify designer preferences. We generalize results from the analysis of animal foraging behavior to model the AAV. Then, using a performance metric common in behavioral ecology, we explicitly find the optimal speed and task processing choice policy for a version of the AAV problem. Finally, in simulation, we show how parameter estimation can be used to determine the optimal controller online when density of task types is unknown.

An Enzyme-Inspired Approach to Stochastic Allocation of Robotic Swarms Around Boundaries

This work presents a novel control approach for allocating a robotic swarm among boundaries. It represents the first step toward developing a methodology for encounter-based swarm allocation that incorporates rigorously characterized spatial effects in the system without requiring analytical expressions for encounter rates. Our approach utilizes a macroscopic model of the swarm population dynamics to design stochastic robot control policies that result in target allocations of robots to the boundaries of regions of different types. The control policies use only local information and have provable guarantees on the collective swarm behavior. We analytically derive the relationship between the stochastic control policies and target allocations for a scenario in which circular robots avoid collisions with each other, bind to boundaries of disk-shaped regions, and command bound robots to unbind. We validate this relationship in simulation and show that it is robust to environmental changes, such as a change in the number or size of robots and disks.

Generalizing foraging theory for analysis and design

Foraging theory has been the inspiration for several decision-making algorithms for task-processing agents facing random environments. As nature selects for foraging behaviors that maximize lifetime calorie gain or minimize starvation probability, engineering designs are favored that maximize returned value (e.g. profit) or minimize the probability of not reaching performance targets. Prior foraging-inspired designs are direct applications of classical optimal foraging theory (OFT). Here, we describe a generalized optimization framework that encompasses the classical OFT model, a popular competitor, and several new models introduced here that are better suited for some task-processing applications in engineering. These new models merge features of rate maximization, efficiency maximization, and risk-sensitive foraging while not sacrificing the intuitive character of classical OFT. However, the central contributions of this paper are analytical and graphical methods for designing decision-making algorithms guaranteed to be optimal within the framework. Thus, we provide a general modeling framework for solitary agent behavior, several new and classic examples that apply to it, and generic methods for design and analysis of optimal task-processing behaviors that fit within the framework. Our results extend the key mathematical features of optimal foraging theory to a wide range of other optimization objectives in biological, anthropological, and technological contexts.

Distributed and Cooperative Task Processing: Cournot Oligopolies on a Graph

This paper introduces a novel framework for the design of distributed agents that must complete externally generated tasks but also can volunteer to process tasks encountered by other agents. To reduce the computational and communication burden of coordination between agents to perfectly balance load around the network, the agents adjust their volunteering propensity asynchronously within a fictitious trading economy. This economy provides incentives for nontrivial levels of volunteering for remote tasks, and thus load is shared. Moreover, the combined effects of diminishing marginal returns and network topology lead to competitive equilibria that have task reallocations that are qualitatively similar to what is expected in a load-balancing system with explicit coordination between nodes. In the paper, topological and algorithmic conditions are given that ensure the existence and uniqueness of a competitive equilibrium. Additionally, a decentralized distributed gradient-ascent algorithm is given that is guaranteed to converge to this equilibrium while not causing any node to over-volunteer beyond its maximum task-processing rate. The framework is applied to an autonomous-air-vehicle example, and connections are drawn to classic studies of the evolution of cooperation in nature.

Self-referencing cellular automata: A model of the evolution of information control in biological systems

Cellular automata have been useful artificial models for exploring how relatively simple rules combined with spatial memory can give rise to complex emergent patterns. Moreover, studying the dynamics of how rules emerge under artificial selection for function has recently become a powerful tool for understanding how evolution can innovate within its genetic rule space. However, conventional cellular automata lack the kind of state feedback that is surely present in natural evolving systems. Each new generation of a population leaves an indelible mark on its environment and thus affects the selective pressures that shape future generations of that population. To model this phenomenon, we have augmented traditional cellular automata with state-dependent feedback. Rather than generating automata executions from an initial condition and a static rule, we introduce mappings which generate iteration rules from the cellular automaton itself. We show that these new automata contain disconnected regions which locally act like conventional automata, thus encapsulating multiple functions into one structure. Consequently, we have provided a new model for processes like cell differentiation. Finally, by studying the size of these regions, we provide additional evidence that the dynamics of self-reference may be critical to understanding the evolution of natural language. In particular, the rules of elementary cellular automata appear to be distributed in the same way as words in the corpus of a natural language.

Physical stigmergy for decentralized constrained optimization: an intelligent lighting example

Conventional distributed solutions for optimization problems with inseparable constraints require significant coordination between agents. Here, a novel numerical approach is described that achieves coordination via stigmergy - agents communicate indirectly through modifications of the environment. This approach is designed for optimal resource allocation problems where every solution exists on a constraint boundary; these boundaries provide the environmental cues that guide the collective motion of the distributed actuators without formal communication between them. Theoretical and experimental results validate this approach for an intelligent lighting example; despite the lack of direct coordination, Pareto-optimal allocations are stabilized. The general approach of using physical stigmergic memory may be useful in many other cyber-physical systems.

Superorganismic Behavior via Human Computation

In a future world with pervasive Human Computation (HC), there may be profound effects on how humanity functions at multiple levels from individual behaviors to species/wide changes in evolutionary development. What would such an HC/shaped human society look like? This hypothetical society would be the result of successful adaptations that provide both increased benefit to the high/level facilitators of large-scale computations as well as sufficient incentives to individuals to participate in those computations.

Utilizing S-TaLiRo as an automatic test generation framework for autonomous vehicles

This paper proposes an approach to automatically generating test cases for testing motion controllers of autonomous vehicular systems. Test scenarios may consist of single or multiple vehicles under test at the same time. Tests are performed in simulation environments. The approach is based on using a robustness metric for evaluating simulation outcomes as a cost function. Initial states and inputs are updated by stochastic optimization methods between the tests for achieving smaller robustness values. The test generation framework has been implemented in the toolbox S-TaLiRo. The proposed frameworks ability to generate interesting test cases is demonstrated by a case study.

The Sunk-cost Effect as an Optimal Rate-maximizing Behavior

Optimal foraging theory has been criticized for underestimating patch exploitation time. However, proper modeling of costs not only answers these criticisms, but it also explains apparently irrational behaviors like the sunk-cost effect. When a forager is sure to experience high initial costs repeatedly, the forager should devote more time to exploitation than searching in order to minimize the accumulation of said costs. Thus, increased recognition or reconnaissance costs lead to increased exploitation times in order to reduce the frequency of future costs, and this result can be used to explain paradoxical human preference for higher costs. In fact, this result also provides an explanation for how continuing a very costly task indefinitely provides the optimal long-term rate of gain; the entry cost of each new task is so great that the forager avoids ever returning to search. In general, apparently irrational decisions may be optimal when considering the lifetime of a forager within a larger system.