Cooperative collision avoidance strategies for unmanned aerial vehicles
| dc.contributor.advisor | Engelbrecht, Japie | en_ZA |
| dc.contributor.author | Meiring, Lauren | en_ZA |
| dc.contributor.other | Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering. | en_ZA |
| dc.date.accessioned | 2021-11-29T05:06:48Z | |
| dc.date.accessioned | 2021-12-22T14:24:51Z | |
| dc.date.available | 2021-11-29T05:06:48Z | |
| dc.date.available | 2021-12-22T14:24:51Z | |
| dc.date.issued | 2021-12 | |
| dc.description | Thesis (PhD)--Stellenbosch University, 2021. | en_ZA |
| dc.description.abstract | ENGLISH ABSTRACT: In this dissertation, a trajectory planning based cooperative collision avoidance system is proposed that provides integrated aircraft-to-aircraft collision avoidance and terrain avoidance for multiple independent UAVs. The UAVs use horizontal, vertical or three- dimensional manoeuvres to avoid short-term collisions with one another, with static ter- rain and with dynamic obstacles while minimising the deviation from their planned long term flight paths. Two existing strategies, namely a centralised strategy and a decoupled strategy, are ap- plied and a novel semi-centralised strategy is developed. The semi-centralised strategy is a hybridisation of the centralised and decoupled strategies. The semi-centralised strategy creates groups of UAVs involved in the same potential collisions and performs centralised trajectory planning for each group. The semi-centralised strategy is further developed into two variants, a semi-centralised expanding strategy and a semi-centralised token-passing strategy. The four cooperative collision avoidance strategies are implemented and tested in a sim- ulation environment. Monte Carlo simulations are performed to evaluate and compare their performances statistically. The trajectory planning strategies are compared using three performance metrics: their success rate at finding solutions, the optimality of the solutions in terms of the chosen cost function, and the time taken to find the solution. The results of the Monte Carlo simulations show that, given enough time, the centralised strategy always finds a solution, if it exists, and finds the optimal solution. However, it takes longer to find a solution. The decoupled strategy finds solutions the fastest. However, the decoupled strategy is not guaranteed to find a solution as its success is dependent on the UAV priority order used. The two semi-centralised strategies have a higher success rate and provide more optimal solutions than the decoupled strategy while finding solutions significantly faster than the centralised approach. | en_ZA |
| dc.description.abstract | AFRIKAANSE OPSOMMING: Die navorsing stel ’n padbeplanning-gebaseerde samewerkende botsingvermydingstelsel voor wat geïntegreerde vliegtuig-tot-vliegtuig botsingvermyding en terreinvermyding vir veelvoudige onafhanklike UAVs verskaf. Die UAVs gebruik horisontale, vertikale of drie- dimensionele maneuvers om korttermyn botsings met mekaar, statiese hindernisse en di- namiese hindernisse te vermy terwyl die UAVs se afwyking van hulle langtermyn vlug- planne geminimeer word. Twee bestaande strategieë word toegepas, naamlik ’n sentrale strategie en ’n ontkoppelde strategie, en ’n nuwe semi-sentrale strategie word voorgestel. Die semi-sentrale strategie is ’n hibriede van die sentrale strategie en die ontkoppelde strategie. Die semi-sentrale strategie skep groepe van UAVs wat betrokke is by dieselfde botsings en doen sentrale padbeplanning vir elke groep. Die semi-sentrale strategie word verder ontwikkel in twee variante, ’n semi-sentrale uitbreidende strategie en ’n semi-sentrale toekenning-aanstuur strategie. Die vier samewerkende botsingsvermydingstrategieë is geïmplimenteer en getoets in ’n simulasie omgewing. Monte Carlo simulasies is uitgevoer om die prestasies van die strate- gieë statisties te evalueer en vergelyk. Die botsingsvermydingstrategieë is evalueer op grond van drie maatstawwe: die sukseskoers, die optimaliteit van die oplossings in terme van die gekose kostefunksie, en die tyd geneem om ’n oplossing te kry. Die resultate van die Monte Carlo simulasies wys dat, gegee genoeg tyd, sal die sentrale strategie die optimale oplossing vind indien dit bestaan, maar neem te lank om ’n oplossing te vind. Die ontkoppelde strategie vind oplossings die vinnigste, maar is nie gewaarborg om ’n oplossing te vind nie, en die sukses is afhanklik van die prioriteitsorde van die UAVs. Die twee semi-sentrale strategieë het ’n hoër sukseskoers en vind meer optimale oplossings as die ontkoppelde strategie en vind oplossings aansienlik vinniger as die sentrale strategie. | af_ZA |
| dc.description.version | Doctoral | en_ZA |
| dc.format.extent | 177 pages | en_ZA |
| dc.identifier.uri | http://hdl.handle.net/10019.1/123847 | |
| dc.language.iso | en_ZA | en_ZA |
| dc.publisher | Stellenbosch : Stellenbosch University | en_ZA |
| dc.rights.holder | Stellenbosch University | en_ZA |
| dc.subject | UCTD | en_ZA |
| dc.subject | UAVs (Unmanned aerial vehicles) -- Collision avoidance | en_ZA |
| dc.subject | Aeronautics -- Safety measures | en_ZA |
| dc.subject | UAVs (Unmanned aerial vehicles) -- Collision avoidance systems | en_ZA |
| dc.title | Cooperative collision avoidance strategies for unmanned aerial vehicles | en_ZA |
| dc.type | Thesis | en_ZA |