http://hdl.handle.net/2268.2/6061 Thu, 05 Mar 2026 22:28:59 GMT 2026-03-05T22:28:59Z http://hdl.handle.net/2268.2/18256 Title: Neuromorphic control of embodied central pattern generators Abstract: The control of robotic locomotion poses important challenges. In particular, we are still very far from achieving robotic locomotion control with the same degree of robustness and adaptability to unexpected environmental perturbations exhibited by moving biological systems. This master’s thesis aims to create a robust and efficient controller for regulating a simple mechanical system. Biological neuron models are used to create artificial central pattern generators (CPGs) that form the core of the controller. Similar to Yu et al., the inspiration of this thesis is the known electrophysiology, sensory response, and modulation of biological CPGs. This study explores the control of a simple resonant mechanical system (a pendulum) to achieve high-amplitude periodic motion without fine-tuning the neuron parameters and with sensory feedback and weak actuation. The design follows multiple steps. It starts with the design and tuning of the controller using a single neuron. This uncovers that only the motor neurons exhibiting a robust type of bursting are able to robustly and easily adapt their excitable behavior to the unknown mechanical system’s properties (damping, resonant frequency, mass, etc.). This is followed by the natural addition of another motor neuron to form a CPG and make the controller symmetric. This increases the achievable amplitude and improves the resilience to perturbations in the controller parameters. Then, neuromodulation is added to allow the dynamic change of the controller properties to control the amplitude of the oscillations. This leads to a trade-off between the speed of convergence to the desired amplitude and the stability of the controller. Finally, multiple controller-pendulum systems are interconnected at the controller level to achieve the desired spatiotemporal pattern between the pendulums. The results indicate that the neuromorphic approach is well-suited for the design of robust controllers. The proposed controller demonstrates the ability to easily adapt to the mechanical system properties to achieve the amplitude goal, as well as the ability to interconnect in a network of controllers. Extensions of the model could be used to control locomotion in robotics or other domains. Sun, 03 Sep 2023 22:00:00 GMT http://hdl.handle.net/2268.2/18256 2023-09-03T22:00:00Z http://hdl.handle.net/2268.2/18255 Title: Data assimilation as simulation-based inference Abstract: Complex dynamical systems are found across various scientific disciplines, representing phenomena like atmospheric and oceanic behavior, brain activity, robot state in its environment, among many others. Due to the challenges that those systems may address, it is often impractical to observe their complete state, leading to the collection of partial observations. For instance, weather stations can only measure a limited number of variables like temperature and pressure, but not the entire state of the atmosphere. However, despite the limited nature of those observations, we can still use them to infer and deduce states that are consistent with the gathered data. By leveraging advanced inference methods, we can make predictions about the complete state of complex dynamical systems based on these information. In this thesis, we delve into the realm of simulation-based inference methods applied to inverse problems in high-dimensional dynamical systems. We discuss how classical methods can be adapted to our problem and investigate how existing evaluation techniques can be used to assess our estimator's performances. Unlike classical simulation-based inference problems, our focus extends to incorporating the temporal dimension of such systems and scaling consistently existing inference methods with the size of the problem. Our goal is to infer the posterior density of states in dynamic systems, using observations to condition the inference process. By accounting for the temporal aspect, we can extend our understanding of the system's behavior and make informed predictions about its future states. Eventually, we show that existing estimation methods can adapt to our problem by incorporating consistently available information related to both system dynamics and observation process. We argue that convolutional estimators are needed to allow good scaling without increasing excessively computational costs. By leveraging system's structure, we found diffusion-based estimators being promising to solve our problem. We also highlight the need of new evaluation techniques that scales correctly and propose a classifier-based posterior check that fill the lacks of other classical evaluations at the cost of harder interpretation. Sun, 03 Sep 2023 22:00:00 GMT http://hdl.handle.net/2268.2/18255 2023-09-03T22:00:00Z http://hdl.handle.net/2268.2/18193 Title: Design and Performance Analysis of a Neuromodulable Spiking Recurrent Cell in the Context of Supervised Learning Abstract: From the desire to understand how neurons work and how we learn, neuroscientists from early 1900 to the 1950s proposed increasingly more complex neuron models. However, no learning rule seemed to have a satisfactory performance until the invention of back-propagation which uses optimisation and calculus to optimise the performance at a certain task. This new learning rule required neural networks to steer away from highly non-linear biological neuron models to ones with continuous properties that allow them to work. Machine Learning researchers favoured this new technique and have obtained increasingly better results with increasingly larger networks, large amounts of data, more computing power necessary to use them, and inevitably, higher power requirements. In an attempt to solve the shortcomings of modern artificial neural networks, the machine learning community is looking again into bio-inspired spiking neural networks and new learning rules that would perform as well as backpropagation. To this date, these attempts have had varying levels of success but have all fallen short of the modern standards of machine learning. However, a recent contribution has proposed a new way of creating and training spiking neural networks by re-designing classical recurrent neural network cells to force them to exhibit spiking behaviour. This allows the use of backpropagation without the need for approximations or adaptations as other spiking neural network training methods have proposed in the past. In this work, we analyse the proposed spiking recurrent cell from a nonlinear systems dynamics perspective and observe some of its shortcomings. Then, we propose two new cells that extend and reduce some of the previously observed issues. Namely, we extend the cell by introducing neuromodulation capabilities to have 3 distinct types of excitability through the addition of a single parameter, and with the second proposed cell, we extend further its capabilities by introducing bursting behaviour. Finally, we perform tests on small neural networks composed of 42 neurons based on the former cell on the MNIST benchmark coded in spikes through latency, and rate coding. In the experiments, we first modulate the neurons to fixed firing types and analyse the performance and behaviour of the network on the 2 tasks. We then randomly initialise the modulation through the network and set it as a learnable parameter for the same tasks. In these tests, we have achieved 91\% accuracy which is close to state-of-the-art performance on SNNs that contain more than 100 times the number of neurons. Furthermore, we have observed that in the current setup, the network does not use a heterogeneity of firing types in the network to obtain the best results. Instead, it converges to one of the newly introduced firing types. Concisely, the main goal of this work is to introduce neuromodulation to a spiking neural network cell that learns through direct backpropagation, and show that modulation can also be learned through this same learning algorithm to obtain satisfactory results. Sun, 03 Sep 2023 22:00:00 GMT http://hdl.handle.net/2268.2/18193 2023-09-03T22:00:00Z http://hdl.handle.net/2268.2/17690 Title: Environmental monitoring at the Montefiore institute thanks to a LoRa network Abstract: In a world where everything becomes automated and digital, the development of connectivity technologies is a key element. In this context, the emergence of wireless communication devices revolutionizes the possibilities that are offered, especially for environmental monitoring. These devices are part of a wider active research topic called Internet Of Things (IoT). This work aligns with this overall context. It aims at creating an autonomous network communicating via a LoRa (long-range) wireless protocol to monitor indoor environmental parameters. In particular, the goal is to create, from scratch, several standalone nodes that sense their environment and transmit the obtained data to a central gateway. This gateway finally shares the obtained data with a network for user display. The thesis begins with the development of a simplified network. This first step aims at developing the main elements that can then be used in the global network. In this first part, several sensors, interfaced with a development board, transmit data via wireless communication to the gateway. This part also includes a complete description of the choices that have been made regarding communication protocols and sensors. The work then focuses on the creation of autonomous nodes. This aspect is related to the autonomous supply voltage of the nodes via indoor solar panels as well as the creation of the node antenna for data transfer. The remaining part of this thesis is dedicated to the description of the nodes. In particular, this section explains how they have been designed and implemented, both from a software and hardware point of view. The final section then describes how the gateway sends data to the created network. Sun, 25 Jun 2023 22:00:00 GMT http://hdl.handle.net/2268.2/17690 2023-06-25T22:00:00Z