Browsing by Author "Hamid, Alala Mohammed Zain Ba"

Number of people with hand disability is increasing recently due to diseases and accidents. Rehabilitation and assistive robotics is an emerging field of research where researchers are trying to develop tailored made robotic devices to address the challenge of disability. However choice of proper actuator for designing assistive device is important to make it safe and compact. Moreover, feedback integrated system of structure-less wearable hand for partially disabled people is not available yet so this dissertation contributes to make it done. This work presents a study on development of two wearable assistive robotic hands for grasping based on Shape Memory Alloy (SMA) and Linear Actuator. The proposed designs are compact and sufficiently light to be used as an assistive hand. Each design is tendon driven and joint-less structure that has the potential to be used as an assistive device for stroke patients. The proposed designs have been implemented for index and thumb fingers as a first prototype to enable grasping. SMA actuator and bias force mechanism were used together in the first design for the purpose of hand's flexion and extension where a linear actuator was used in the second design. This work provides a review on the previous work done in the field of hand rehabilitations and describes the mechatronic design of the wearable hand, simulation, modeling, and development of the actuation force mechanism and sensory system. Experiments of open loop controller were conducted to understand the hand characterization and grip force provided by index finger. A feedback controller (proportional controller) has been implemented for this prototype with gripping force as the feedback parameter. In the first design, it was observed that approximately 2.25 A current caused 4 cm displacement for SMA actuator. The maximum temperature of the SMA actuator was achieved to be 100 Celsius. The attainable gripping force was around 2N for a load free finger. The conducted experiments showed promising results that encourages further development on this. The same procedures were implemented in the case of Linear Actuator. However Linear Actuator consumed much less power in comparison with SMA. Performance comparison and evaluation between the assistive hands based on SMA and Linear Actuator were conducted. Various experiments were carried out to explore the performance behavior of these two actuators.

Understanding human response to crowd emergency is extremely complex and it plays a significant role in engineering construction designs and crowd safety. Individual choices, reasoning, and behaviours can’t be fully described by equations or rule-based methods. Accordingly, this research proposes a neuro-symbolic approach for modelling agents with human-level capabilities of reasoning and performance in an emergency evacuation. The proposed neuro-symbolic approach combines deep reinforcement learning (DRL) with evaluative fuzzy logic function to address the challenges of large amounts of required data, time, and trials-and-errors for policy optimization and to handle the assumption of the reward functions that may not be practical in real scenarios. This neuro-symbolic model has the potential to deal with the complexity of the environment and decision-making process via DRL and enhances the cognitive and visual intelligence via an evaluative fuzzy function, which continuously evaluates agent actions during the training process to boost pedestrian active response to their surroundings, with full awareness of time, thereby, a human-level capacity of reasoning. Moreover, this proposed model optimizes the computational demands of DRL and enables faster learning of new situations. A crowd disaster can easily occur all around. Thereby, prediction of the critical conditions is significant to prevent terrible disasters. Earlier research on entropy-based prediction models was primarily based on either the density or velocity attributes. Additionally, they evaluated the crowd risk by determining the level of chaos or abnormal behaviour rather than predicting a crowd disaster. Therefore, this research also proposes a crowd Boltzmann entropy-based prediction model that integrates the local density with average local speed to identify the critical situations that may result in crowd disasters, empowering sufficient prediction of crowd critical conditions and accurate description of the nature of a crowd motion therefore enabling early preventative intervention. The findings indicate that the proposed neuro-symbolic model can produce behavioral patterns that align with real observations of crowd evacuation, such as laminar flow, stop-and-go flow, and crowd turbulence. On top of that, a new evacuation behavior is observed, as some pedestrians avoid congestion at the exit until the density reduces. Moreover, the proposed model illustrates a higher accuracy and much faster converge than the pure PPO model with substantially minimal training timesteps as little as 2 to 8 percentage. Results of the proposed entropy-based prediction model reveal that the critical conditions last the longest in the nearest area to exit with the highest average entropy 0.44, and the shortest in the farthest area to exit with the lowest average entropy of 0.29 which is consistent with crowd literature. Meanwhile, the reliability study records an increase of the mean and standard deviation of evacuation time from 39.7s, 1.06 to 155.09s, 7.39 as crowd size increases from 15 to 200 pedestrians which implies a rise of uncertainty. The proposed prediction model outperforms well-established prediction approaches such as crowd pressure and crowd flow. Therefore, this work can provide crowd authorities and construction engineers with insights into complex crowd behaviour and critical conditions to evaluate evacuation plans and make decisions and sustainable designs to ensure crowd safety.