Ludvig Ericson

About

I am a researcher at Ericsson Research, where I work on cyber-physical systems — in particular localization, RF environmental reconstruction, and integrated sensing and communication (ISAC). The common thread is teaching radios to perceive the world they transmit through, and using what they learn to plan, map, and act.

I came to this from robotics. In 2025 I defended my PhD at KTH Royal Institute of Technology, in the Robotics, Perception and Learning division, under the supervision of Patric Jensfelt. My thesis, Exploration and Prediction: Beyond-the-Frontier Autonomous Exploration in Indoor Environments, studies how a robot can use learned priors over indoor floor plans to plan more efficiently in partially observed worlds.

Outside the day job, I run Send a Patch, a small Swedish software firm I founded as a student, and I have been writing open-source Python for over a decade — pylibmc, simples3, and fish are still in active use across the ecosystem.

Research

My work sits at the intersection of mobile robotics, SLAM, and learned spatial models. The papers below are listed in reverse chronological order; a full list and citation metrics are on Google Scholar.

• 2025
  Information Gain Is Not All You Need

  Ludvig Ericson, José Pedro, Patric Jensfelt

  European Conference on Mobile Robots (ECMR)

  Project PDF arXiv Code Videos
• 2025
  Exploration and Prediction: Beyond-the-Frontier Autonomous Exploration in Indoor Environments

  Ludvig Ericson

  PhD thesis, KTH Royal Institute of Technology

  PDF DiVA
• 2024
  Beyond the Frontier: Predicting Unseen Walls from Occupancy Grids by Learning from Floor Plans

  Ludvig Ericson, Patric Jensfelt

  IEEE Robotics and Automation Letters, vol. 9, no. 8

  Project PDF arXiv Demo Dataset
• 2024
  ExelMap: Explainable Element-based HD-Map Change Detection and Update

  Lena Wild, Ludvig Ericson, Rafael Valencia, Patric Jensfelt

  Preprint, arXiv:2409.10178

  arXiv
• 2022
  FloorGenT: Generative Vector Graphic Model of Floor Plans for Robotics

  Ludvig Ericson, Patric Jensfelt

  IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

  Project PDF arXiv Video Code Dataset
• 2021
  Understanding Greediness in Map-Predictive Exploration Planning

  Ludvig Ericson, Daniel Duberg, Patric Jensfelt

  European Conference on Mobile Robots (ECMR)

  PDF IEEE Code
• 2019
  GCNv2: Efficient Correspondence Prediction for Real-Time SLAM

  Jiexiong Tang, Ludvig Ericson, John Folkesson, Patric Jensfelt

  IEEE Robotics and Automation Letters, vol. 4, no. 4

  arXiv IEEE Code
• 2017
  On the Performance of Network Parallel Training in Artificial Neural Networks

  Ludvig Ericson, Rendani Mbuvha

  Preprint, arXiv:1701.05130

  arXiv
• 2017
  On the Use of Unmanned Aerial Vehicles for Autonomous Object Modeling

  Michael C. Welle, Ludvig Ericson, Rares Ambrus, Patric Jensfelt

  European Conference on Mobile Robots (ECMR)

  DiVA

Software

A selection of open-source projects. The complete list lives on GitHub.

• pylibmc

  C / Python · 493★

  A Python wrapper around the libmemcached interface from TangentOrg. Long-running maintainer.

• simples3

  Python · 93★

  A small, dependency-free interface to Amazon S3 from Python.

• fish

  Python · 68★

  The Python module that swims — lightweight progress bars and animations for the terminal.

• git-remote-deploy

  Shell · 22★

  Automatic deploy on git push, using only Git's native remote-helper protocol.

• procname

  C extension · 15★

  Set the argv[0] process title from inside Python programs.

• pysquad

  Python / C++

  High-performance simulation of quadrotors and other multicopters.

• floorgent

  Research code

  Reference implementation accompanying the FloorGenT paper (IROS 2022).

• da

  Research code

  Reproduction of Distance Advantage, the exploration heuristic from the ECMR 2025 paper.

• pam-watchid

  Swift

  A PAM module that lets the Apple Watch authenticate sudo on macOS.

• torch-occupancy-grid-bfs

  PyTorch

  Vectorised breadth-first search across batches of occupancy grids, expressed as 2D convolutions.

Other projects

Smaller things I've made over the years and still recommend.

• Coders Dvorak — a Dvorak-derived keyboard layout for macOS, optimised for programming punctuation.
• Pokemon Names with Deep Learning — lightning talk and slides from EuroPython 2016.
• Cosmonaut — an interactive Newtonian-gravity sandbox in the browser.
• GPSpeed — an iOS speedometer using only GPS.

Contact

Email

ludvig at@ lericson.se

GitHub

github.com/lericson

Scholar

Google Scholar

Company

sendapatch.se