Virginia Commonwealth University VCU Scholars Compass Theses and Dissertations Graduate School 2013 Actual Entities: A Control Method for Unmanned Aerial Vehicles Erica Absetz Virginia Commonwealth University Follow this and additional works at: https://scholarscompass.edu/etd Part of the Computer Sciences Commons © The Author Downloaded from https://scholarscompass.edu/etd/3094 This Thesis is brought to you for free and open access by the Graduate School at VCU Scholars Compass. It has been accepted for inclusion in Theses and Dissertations by an authorized administrator of VCU Scholars Compass. For more information, please contact libcompass@vcu. Actual Entities: A Control Method for Unmanned Aerial Vehicles A thesis submitted in partial fulfillment of the requirements for the degree Master of Science at Virginia Commonwealth University by Erica Lynn Absetz Bachelor of Science, Virginia Commonwealth University, 2012 Associate of Science, J.
Sergeant Reynolds Community College, 2009 Director: Dr. James Ames Associate Professor, Department of Computer Science Virginia Commonwealth University Richmond, VA April, 2013 ii © Erica L. Absetz 2013 All Rights Reserved iii Acknowledgements There have been many people that have supported me at different times in my education that I would like to thank. Thank you to both of my parents for their undying love and support, it motivated me to do the best I could throughout my educational career.
Thank you to my best friend, Johnny Xmas, for threatening to never speak to me again if I didn’t finish my degree. That helped me power through the rough patches. I would also like to thank Dr. Ames as he took on the last minute task of advising my thesis for this last semester.
Above all, I would like to acknowledge the late Dr. Primeaux and the impact that he has had on my educational path. His passion and energy made me fall in love with Artificial Intelligence. He encouraged my pursuit of a Master’s Degree and was a valued mentor for the beginning of this thesis as he guided the formation of the concept.
iv Table of Contents Copyright.iii List of Figures.13 7 Colony of Prehending Entities.14 8 Example Situation of Actual Entities.17 11 Centralized/Decentralized Planners.1 Control Strategies – Baseline.2 Control Strategies – Random.3 Control Strategies – Repulsion.4 Control Strategies – Pheromones.5 Control Strategies – Evolutionary Algorithms.24 13 Actual Entity Problem Domain.1 Simulation of the UAVs and Search Area.28 14 Actual Entity Behavioral Control.1 Actual Entity Representation.2 Eternal Object Representation.32 15 Design of Experiments.1 Results and Statistical Analysis.43 Appendix A: Environment Simulator Class Code.44 Appendix B: COPE Class Code.47 Appendix C: Actual Entity Class Code.50 Appendix D: Eternal Object Class Code.65 v List of Figures: Fig 1: Inner and Outer Thresholds.12 Fig 2: Baseline Efficiency.21 Fig 3: Simulation of Repulsion Control Strategy.22 Fig 4: Simulation of Pheromone Control Strategy.23 Fig 5: Path Planning with Digital Pheromones.24 Fig 6: Genetic Algorithm - Crossover.25 Fig 7: Chromosome of Behavior Archetype Genes.26 Fig 8: Overview of an Evolutionary Algorithm Method.27 Fig 9: Beginning of Simulation – Target Coordinates.33 Fig 10: Beginning of Simulation for 10 UAVs and 5 Targets.34 Fig 11: End of Simulation.35 Fig 12: Number of Searches Completed vs. Convergence Time Constraint.36 Fig 13: Number of Searches Completed at Time Constraints Separated by Thresholds.37 Fig 14: Number of Completed Searches Based on Number of UAVs.38 Fig 15: Number of Targets Compared to Searches Successfully Completed.40 Fig 16: Emergent Search Pattern with 10 UAVs.40 Fig 17: Emergent Search Pattern with 5 UAVs.41 Abstract ACTUAL ENTITIES: A CONTROL METHOD FOR UNMANNED AERIAL VEHICLES By Erica Lynn Absetz, M. A thesis submitted in partial fulfillment of the requirements for the degree Master of Science at Virginia Commonwealth University. Virginia Commonwealth University, 2013 Director: Dr.
James Ames, Associate Professor, Department of Computer Science The focus of this thesis is on Actual Entities, a concept created by the philosopher Alfred North Whitehead, and how the concept can be applied to Unmanned Aerial Vehicles as a behavioral control method. Actual Entities are vector based, atomic units that use a method called prehension to observe their environment and react with various actions. When combining multiple Actual Entities a Colony of Prehending Entities is created; when observing their prehensions an intelligent behavior emerges. By applying the characteristics of Actual Entities to Unmanned Aerial Vehicles, specifically in a situation where they are searching for targets, this emergent, intelligent behavior can be seen as they search a designated area and locate specified targets.
They will alter their movements based on the prehensions of the environment, surrounding Unmanned Aerial Vehicles, and targets. Chapter 1: Introduction A philosopher, Alfred North Whitehead, created the concept of Actual Entities. An actual entity is an atomic unit that changed based on its prehensions of the surrounding environment. Prehension, a term that Whitehead created, is defined as an observation followed by reactions.
When multiple Actual Entities are combined into the same environment a colony of prehending entities is created, which exhibits an emerging, intelligent behavior. The use of Actual Entities had been introduced into the field of computer science by both Primeaux and Saunders [8, 9, 11]. They have used the concept as a modeling tool, a way to find global and local maxima in a 3D space, and as a trust-based learning method for data characteristics. Both acknowledge Actual Entities to be a form of Artificial Intelligence belonging to the classification of Swarm Intelligence (SI).
SI is commonly used for most control methods for Unmanned Aerial Vehicles. Now, take a look at Unmanned Aerial Vehicles and how they are sent out in groups to complete a mission. There have been many different types of behavioral control methods created to find the most efficient and dynamic control method. It is important that the control method can work with changing numbers of Unmanned Aerial Vehicles and to not rely heavily on communications.
This thesis will be used to show that Actual Entities can be used as a behavioral control method for Unmanned Aerial Vehicles. It will focus primarily on the application of Actual Entities to Unmanned Aerial Vehicles to complete a searching objective of locating designated targets. Furthermore, the concept of Actual Entities and that of other control methods currently in use will be covered. Finally, the control method will be tested in a simple simulation and the results will be analyzed.
Coding samples used for the testing of this concept will be provided in the appendix. 7 Chapter 2: Actual Entities Actual entities (AE) are the atomic units of the theory of a philosopher by the name of Alfred North Whitehead. As Hooper [4] states, “actual entities compose the Universe of existence in space and time, and it is therefore important first of all to understand what these entities are.” AEs are vector based, a behavior involves an AE and an object then defines a vector relationship between the two. It is important to keep in mind that an AE’s current state is a result of its history and the two cannot be separated.
An actual entity is essentially a “process”, containing its growth between phases, ending in a final and definite achievement. Hooper [4] stated that AEs could be referred to as genetic in that they have a process to “become” and the outcome of that causes “satisfaction”. Every AE has its own unique path in its development and it uses its own perspective on the world to mirror it. As discussed by Hooper, actual entities have two sides.
One is the emergent creature formed from its achievements and the other is a mode of creation that uses its own acts to gain experience while pulling the Universe together. A passage from one of Whitehead’s books [12] helps tie it all together as follows: An event has contemporaries – This means that an event mirrors within itself the modes of its contemporaries as a display of immediate achievement. An event has a past. This means that an event mirrors within itself the modes of its predecessors as memories which are fused into its own present.
An event has a future. This means that an event mirrors within itself such aspects as the future throws back into the present, or in other words, as the present has determined concerning the future. Thus an event has anticipation: “The prophetic soul Of the wide world dreaming of things to come.” 8 Chapter 3: Eternal Objects Actual entities are not the only class of entities that exist in the Universe as described by Whitehead, there is another class called eternal objects. These eternal objects are similar to an actual entity except they do not prehend, as they do not exist in the same manner that actual entities do.
Hooper [4] reminds us that eternal objects are constant across all environments; therefore, they provide the important function of being a form of definiteness to actual entities. Anything in the Universe that does not have a reference to a definite object in the temporal world is an eternal object [12]. Eternal objects can be prehended by actual entities, which can affect the “experience” of the actual entity but not that of the eternal object. In the case of UAVs, where the UAV would be the AE, there are multiple items that could be considered to be eternal objects.
The simulation of the concept in this paper considers the targets that are being searched for to be eternal objects, as the targets can be prehended; however, they do not prehend or change themselves. Other possible eternal objects in this case could be terrain, such as mountains and trees. 9 Chapter 4: Prehension Prehension is the method in which actual entities change and move. As described by Whitehead [12], the manner in which an actual entity is capable of altering itself is based on its environment.
The definition of the word “prehension” is a datum prehended by the substrate activity present in the emergence of every actual entity [4]. His doctrine of prehensions describes this action of settled occasions in the world being used to form actual entities experiences. According to Whitehead [12] an important part of prehension is the concept of causality and both efficient causation and final causation must be acknowledged. The belief that past events decide the course of future events is efficient causation.
While, final causation is when the “end” is said to play a part in creating the wanted result. Actual entities serve as an example of both types of causation [4]. The effect of the prehension is based on its subjective aim. The guidelines provided are that a prehension occurs between a prehending and prehended object, of which only the prehending object is altered.
There are two types of prehension: positive and negative. A positive prehension prehends and changes the prehending object. Whitehead described positive prehensions as “feelings”, even though they are essentially operations. To get a feeling from an actual entity it has been referred to as “physical” and to get one from an eternal object it is “conceptual” [4].
A negative prehension prehends and does not change the prehending object although it does have an impact on the overall satisfaction of the actual entity as it does not help to improve it. Hooper painted the following illustration of a negative prehension to make the point clear. Just as the refusal by an individual to love things worthy of being loved, or to see the truth of a situation, even though it stares him in the face, affects his character and his general “tone”, so negative prehensions determine in part the emotional tone of the “satisfaction” of the actual entity [4]. Prehensions involving an actual entity prehending another actual entity are 10 more specifically referred as physical prehensions; where in the case of an actual entity prehending an eternal object it is referred to as a conceptual prehension.
Whitehead never went into any detail of how this should be done computationally; however, Primeaux addressed this in [9] among other things.