★ The internship will take place at the FOX team of CRIStAL laboratory, at the University of Lille.
Summary: In this project, we will perform time series analysis in general. The following are the main tasks of time series analysis :
Time series classification: The time series classification is an important topic in time series analysis, and it has been applied in many areas, such as finance, biometrics, networking, and artificial intelligence. Given a time series sequence p, the goal of classification is to classify each of the p time series into a predefined class. In this process, one of the significant steps is to calculate the distance between each of the p time series and other time series in dataset D. Hence, in the field of time series data mining, a distance function or similarity measure is often used to describe the relationship between two time series.
The problem of time series classification has been intensively investigated during the last few years, with the increase in the computational capabilities of computers. It is particularly important to be able to find the distance between time series. We can find many distance measures for the similarity of time series data. Nevertheless, the simple method combining the nearest neighbor (1NN) classifier and some form of dynamic time warping (DTW) distance has been shown to be one of the best-performing time series classification techniques. The DTW algorithm is a classical and well-established distance measure well-suited to the task of comparing time series.
Previously, we performed an experimental study of many such distance computation algorithms for time series analysis. These are dynamic programming-based techniques, like DTW and its variants. Along with that, we also incorporated other dynamic time programming-based methods (not DTW-based) in this experimental study. The study was performed to match the images of scanned words (a domain named Word Spotting). In this project, we would like to extend this experimental study in the context of 1D time series analysis. In this context, we will perform the following work :
- Perform a comparative experimental study of all these algorithms on 1D time series dataset
- For this experimental study, we will use UCR time series dataset. There are 128 different time series data sets in this archive.
Time series analysis using machine learning: For time series classification, it is shown in the literature that Convolution Neural Network (CNN) has outperformed other non machine learning based approaches for capturing the relevant features in time series. In CNN, the feature extraction consists in finding linear combinations between consecutive time steps of a fixed size. The deeper the model is, the more it increases its receptive field. This represents the input space that a point in a certain depth of the network depends on. The larger the receptive field is, the more beneficial it is for the model. In this context, we will work in the following directions :
- We will explore more into details, the various deep learning based models for time series analysis
The following are the principal steps to be followed in this internship :
Collect & Prepare Data :
- Collect raw time-series from sensors (e.g., accelerometer, temperature, audio).
- Preprocess (normalization, windowing, segmentation) to convert signals into usable training samples.
Train a Deep Learning Model: We will use standard DL frameworks (TensorFlow, PyTorch, etc.). The following are the typical models that we will explore :
- 1D Convolutional Neural Network (CNNs): Good for pattern recognition in fixed-length windows
- Temporal Convolutional Neural Network (TCNs): Efficient alternatives to RNNs for sequence tasks.
- Quantized RNNs / LSTM / GRU (lightweight) : For sequential dependencies
Optimize on Edge: Since microcontrollers have limited RAM, flash memory, and computing power :
- Model Quantization — reduce precision (e.g., 32-bit → 8-bit).
- Pruning — remove redundant weights.
- Knowledge Distillation — A smaller model learns from a larger one.
- Architecture Search/Optimization like TinyNAS.
Through such optimization, we wish to dramatically shrink model size and compute needs.
Deploy & Run Inference on MCU:
- Convert the model to a tiny runtime format using the tools/frameworks below.
- The MCU performs inference locally when a new time series arrives.
Tools & Frameworks for TinyML on MCUs :
| Tool | Purpose |
| TensorFlow Lite for Microcontrollers (TFLM) | Run optimized ML models on MCUs with minimal runtime. |
| Edge Impulse | End-to-end platform for data collection, training, optimization, and generating deployable code. |
| MicroTVM (Apache TVM) | Compiler/optimizer generating efficient C code for microcontrollers. |
| CMSIS-NN | Optimized neural network kernels for ARM Cortex-M cores. |
These make it feasible to execute inference within the memory and speed limitations of MCUs.
Tools & Frameworks for TinyML on MCUs :
| Tool | Purpose |
| ESP32 | * Popular low-cost MCU with Wi-Fi/Bluetooth. * Often runs TinyML models with TFLM or Edge Impulse. * Some variants (e.g., ESP32-S3) include vector instructions for AI acceleration. |
| ARM Cortex-M Series MCUs (e.g., STM32) | * Based on full ARM Cortex-M cores — M4/M7/ M33 etc. with optional FPU. * Supported by STM32Cube.AI for optimizing NN models on STM32. * Widely used in TinyML because of good ecosystem support. |
| Arduino Nano 33 BLE Sense | * Affordable development board with onboard sensors. * Built-in support for TensorFlow Lite Micro and ideal for time-series sensor data. |
| Raspberry Pi Pico (RP2040) | * Dual Cortex-M0+ MCU at very low cost. * Capable of TinyML with tools like TFLM and Edge Impulse. |
| Other Specialized Options : Ambiq Apollo3 (SparkFun Edge) | * Ultra-low-power TinyML board. |
| Other Specialized Options : RISC-V based MCUs with accelerators (Kendryte K210) | * Includes neural accelerators for heavier workloads. |
Note : These MCUs are generally suited for inference, not training models locally. The model is trained offline and then deployed to the MCU with optimized weights.
Desired Profile :
- Final-year Master’s student (M2) or engineering student specializing in machine learning, data analysis, or a related field.
- Knowledge of machine learning and deep learning.
- Knowledge and interests in edge devices such as microcontrollers
- Strong mathematical background
- Programming skills (Python).
- Autonomy, rigor, and critical thinking skills.
★ The internship will take place at the FOX team of CRIStAL laboratory, at University of Lille.
Address of the Internship :
CAMPUS Haute-Borne CNRS IRCICA-IRI-RMN
Parc Scientifique de la Haute Borne, 50 Avenue Halley, BP 70478, 59658 Villeneuve d’Ascq Cédex, France.
Candidature :
If this proposal interests you, please send the following documents to Dr. Tanmoy MONDAL (tanmoy.mondal@univ-lille.fr); Redha KASSI (Redha.Kassi@univ-lille.fr), Damien MARCHAL (damien.marchal@univ-lille.fr), Hedia MARZOUKI (hedia.marzouki@univ-lille.fr)
- CV
- Motivation Letter
- Transcripts of grades obtained in Bachelor’s/Master’s/Engineering school as well as class ranking
- Name and contact details of at least one reference person who can be contacted if necessary
