Search

My thesis explores the cultural, social and management context of digital transformation for African carriers, and discusses the elements of social-cultural barriers, obstacles of digital adoption, innovation, organisation change, ecosystem and their influences. Existing studies are mainly based on the presumed conditions in developed markets, including the free market, rule by law, human capital without considering the social and historical obstacles, absorptive capacities of organisations and the influences of foreign ICT enablers. Using a qualitative research paradigm privileges the insiders' perspective. I have examined the different backgrounds of digital transformation in Africa. Path dependence is a significant negative consequence of post-colonial social net and western knowledge dominant environment. Vested interest plus corruption slowed down the process of digital adoption while arbitrary administration causes unnecessary concerns for participants. Traditionalism of Africa is a trade-off between customs of tribes and influence of colonial power. The priority for carriers is reshaping the organisation to take advantage of existing strategic assets, while the improvement of value creation efficiency is carried out by activating the ability of individuals. The new value net breaks the previous development model of low interdependence, maximising the use of external resources based on complementary advantages and sharing. The value net inherits the advantages of flexibility, innovation, quick response, and risk reduction advantages of organisations. Meanwhile, it has 4 following unique characteristics in Africa: Extremely Various Needs of Customers, Cross-industry, Locally Adopted Business Model and High Sensitivity in Costs. Four aspects are examined in value proposition, core competence, incubation and co-value creation to enable value net synergies. My research contributes to the theory of the digital transformation in undeveloped countries, in particular how social relationships and cultural norms are inextricably linked with insiders' understandings of digital transformation. It also contributes to emerging debate about digital maturity research paradigms and methodologies.

Advances in mobility technology are expected to greatly improve our access to transportation and the convenience of life. Many challenges in achieving the safety, efficiency, and comfort of mobility systems still exist. This dissertation leverages advances in modeling and control to highlight several solutions to these challenges. In particular, for safety-critical applications, it is important to guarantee that the system can operate safely anytime even with the occurrence of failures; hence there is a need for systematic approaches to achieving fault-tolerant control (FTC). Fault tolerance must be ensured in multiple ways: for failure detection and isolation (FDI), system reconfiguration/recovery, and safe operation in all normal and failure modes. Meanwhile, the system availability should be maximized instead of being compromised. In this dissertation, this problem is approached through set-theoretic methods by exploiting nesting between constraint admissible, recoverable, and isolatable sets to satisfy state and control constraints in all phases of failure detection, isolation, reconfiguration, and transition to operation in failure mode. The proposed FTC framework integrates a reference governor for safe reference tracking, and the framework can handle stuck/jammed actuators that are common in mobility applications. An aircraft longitudinal flight control example is used for demonstration. The second advance in this dissertation is an eco-driving approach for connected and autonomous vehicles (CAVs) that accounts for the interactive vehicle behavior in traffic. By exploiting and fusing game theory, reinforcement learning (RL), and battery electric vehicle (BEV) powertrain modeling and control, significant opportunities for energy efficiency improvements are highlighted. Specifically, a level-k game-theoretic traffic model is considered to capture the interactions between the ego vehicle and the surrounding human-driven vehicles. Also, a model that represents a BEV is established, and a control policy for operating the vehicle in traffic is trained by RL applied to the reward function that reflects energy efficiency, travel time efficiency, safety, and comfort. The control policy can command both longitudinal and lateral actions, including acceleration, braking, and lane changes, to enhance the possibilities for energy saving. Finally, the third advance in this dissertation is the development of a trajectory optimizer for multistage hybrid powertrain capable of handling both continuous (engine and motor torques) and discrete (engine on/off or gear selection) degrees of freedom. For driving scenarios prioritizing drivability, the shortening of acceleration duration is often limited by jerk since sudden changes of acceleration can cause discomfort to passengers. Specifically, a 40 to 70 kph acceleration event is considered to simulate a passing maneuver for a multistage hybrid electric vehicle (HEV). Case studies with different jerk limits are set up to investigate the trade-off between acceleration duration and jerk, and the acceleration performance is optimized by solving a constrained trajectory optimization problem formulated as a parallelizable multistage mixed-integer problem. The proposed approach, therefore, addresses a complex optimal control problem in the area of vehicle drivability/rideability that is critical for customer acceptance of energy-efficient and autonomous vehicles. The developed methodology for efficient numerical trajectory optimization with continuous and integer variables can be applied to other engineering domains. Overall, this dissertation highlights opportunities for applications of game-theoretic, set-theoretic and optimal control towards the development of safer, more efficient and comfortable future mobility systems. The contributions address the gaps in safety-guaranteed FTC, safe and energy-aware CAV control in shared traffic, as well as in unbiased and easy-to-implement drivability analysis processes.