Software Engineer Intern - Real-time optimisation of robot schedules - Python / Go
Stanley Robotics
Description de l'offre
Context :
At Stanley Robotics, we build a robotized parking solution that automatically stores cars, using a fleet of robots.
In order to increase the efficiency of our solution, we want to optimize the algorithms of the Fleet Management System (FMS), the service in charge of allocating tasks to robots.
Your mission :
In our system, missions are created on the flow to store and retrieve vehicles, and these missions are allocated each time a robot is available, using matching algorithms.
We propose with this internship to explore new approaches, in particular build a complete schedule of all available missions instead of single task allocation, and quantify the performance improvements and losses.
Your mission will be to model this new problem, to design an algorithm to build a complete schedule for each robot, and adapt it in real-time to operational events. Motivation of such algorithms is the optimisation of resources use and reduction of customer delays. You will validate your algorithms using simulation and real data.
You will join our R&D team of experienced developers from various backgrounds, in a challenging, creative and friendly atmosphere.
Profil recherché
Your profile :
-
You are in your last year of engineering school - min 6 months internship, specialized in applied mathematics / operations research
-
Good algorithmic skills
-
Good programming skills in Python and/or Go is a plus
-
Knowledge about metaheuristic algorithms and robustness theory
-
Be able to work autonomously on challenging tasks
-
French and English at a professional level
Context :
At Stanley Robotics, we build a robotized parking solution that automatically stores cars, using a fleet of robots. In order to maximize the efficiency of our product, we need robots that can navigate smoothly in the yard, including in narrow areas, while surrounded by vehicles.
Your mission :
Join the Robotsoft team at Stanley Robotics, where you will work alongside our R&D team of experienced developers from various backgrounds, in a challenging, creative and friendly atmosphere. You will select one of the two key tracks to advance the field of robust robot navigation, with the shared objective of ensuring obstacle avoidance in narrow lanes while accounting for the robot’s operational and physical constraints.
-
Track 1: Model-Free Navigation with Reinforcement Learning (RL):
You will implement and test new motion planning and control algorithms using RL techniques tailored for real-world applications. -
Track 2: Model Predictive Control (MPC) for Motion Planning and Control:
You will design and implement MPC-based navigation algorithms that account for the system’s limitations while operating within tight runtime constraints.
Responsibilities:
-
Design, implement, and test motion planning and control algorithms in C++/Python
-
Focus on obstacle avoidance in constrained environments, such as narrow lanes, while respecting the robot’s limitations (e.g., speed, acceleration, steering angle, etc.)
-
Explore and implement state-of-the-art methods for robot navigation, including MPC- or RL-based techniques
-
Document and present findings, including testing results and performance analysis