Northern Robotics Laboratory

Lidar Scan Registration Robust to Extreme Motions

2022-01-04T00:00:00-05:00

Registration algorithms, such as Iterative Closest Point (ICP), have proven effective in mobile robot localization algorithms over the last decades. However, they are susceptible to failure when a robot sustains extreme velocities and accelerations. For example, this kind of motion can happen after a collision, causing a point cloud to be heavily skewed. While point cloud de-skewing methods have been explored in the past to increase localization and mapping accuracy, these methods still rely on highly accurate odometry systems or ideal navigation conditions. In this paper, we present a method taking into account the remaining motion uncertainties of the trajectory used to de-skew a point cloud along with the environment geometry to increase the robustness of current registration algorithms. We compare our method to three other solutions in a test bench producing 3D maps with peak accelerations of 200 m/s² and 800 rad/s². In these extreme scenarios, we demonstrate that our method decreases the error by 9.26 % in translation and by 21.84 % in rotation. The proposed method is generic enough to be integrated to many variants of weighted ICP without adaptation and supports localization robustness in harsher terrains.

Contributions

A novel motion uncertainty-based point weighting model, allowing the robustness of registration algorithms to be improved when subject to high velocities and accelerations, requiring neither high-fidelity pose estimation nor prior knowledge of the environment.
Experimental comparison of our proposed approach with state-of-the-art approaches on a dataset gathered at high velocities and accelerations.
A highlight of the impact of point cloud skewing on the accuracy of localization and mapping.

Results in Images

The top figure shows the experimental platform used for this work. A lidar and an IMU are mounted on a sensing platform, which is protected by a cage. A rope is attached to the platform and pulled on by an operator to induce high velocities and accelerations during lidar scans. Using this platform, we acquired scans at linear speeds and accelerations up to 3.5 m/s and 200 m/s² and angular speeds and accelerations up to 11 rad/s and 800 rad/s² in an indoor garage. When an operator pulled the rope, we recorded scans with our platform and used our method to build a map of the environment and localize within it. No prior knowledge of the environment is required by our method.

The bottom figure shows a top view of three 3D reconstructions of the environment built while applying peak accelerations up to 194.98 m/s² on our sensing platform. All three maps were built using the same data, with our method only used for the rightmost map. Points in dark purple have high error and points in green have low error. The error was computed against a ground truth map built while moving slowly the sensors. De-skewing significantly impacts the mean error by reducing it from 21.95 cm to 7.27 cm. Further improvement is gained by using our method, leading to a mean error of 6.86 cm. For example, one can observe an improvement in accuracy on the left wall.

Reference

Deschênes, S.-P., Baril, D., Kubelka, V., Giguère, P., & Pomerleau, F. (2021). Lidar Scan Registration Robust to Extreme Motions. 2021 18th Conference on Robots and Vision (CRV). https://doi.org/10.1109/CRV52889.2021.00014

Accurate outdoor ground truth based on total stations

2021-12-13T00:00:00-05:00

In robotics, accurate ground-truth position fostered the development of mapping and localization algorithms through creation of cornerstone datasets. In outdoor environments and over long distance, total stations are the most accurate and precise measurement instruments for this purpose. Most total station- based systems in the literature are limited to three Degrees Of Freedoms (DOFs), due to the use of a single-prism tracking ap- proach. In this paper, we present preliminary work on measuring a full pose of a vehicle, bringing the referencing system to six DOFs. Three total stations are used to track in real time three prisms attached to a target platform. We describe the structure of the referencing system and the protocol for acquiring the ground truth with it. We evaluated its precision in a variety of different outdoor environments, ranging from open-sky to forest trails, and compare it with another popular source of reference position, the Real Time Kinematics (RTK) positioning solution. Results show that our approach is the most precise one, reaching an average positional error of 10 mm and 0.6 deg. This difference in performance was particularly stark in environments where Global Navigation Satellite System (GNSS) signals can be weaker due to overreaching vegetation.

Contributions

A tracking system using multiple total stations to obtain the 6DOFs of a mobile robot in outdoor environment;
A calibration procedure of multiple total stations for a vehicle continuous tracking;
A time synchronization and a long-range data transmis- sion protocol for online pose referencing;
A precision analysis of our tracking procedure; and
A comparison against a RTK positioning system.

Results in Images

To ground the performance of the total station referencing system to a commonly-known alternative, we compared it to the RTK solution. The satellite localization performed well in open areas, but a decreased in accuracy can be observed when portions of the sky view were obstructed by vegetation. To this effect, we have recorded the RTK positioning data during two experiments: one in the quarry and another one in the forest. Similarly to using the inter-prism distance to estimate the total station localization precision, we rely on the measured distance between the GNSS receivers to estimate their precision.

In the top figure, we compare the precision of the corrected RTK localization and our total station referencing system in the quarry. In this environment and with the robot moving, the precision is equivalent as shown on the histogram.

In a more difficult scenario, the bottom figure shows portions of the forest experiment, where the robot was in reach of all three total stations. From the beginning until the mark of 200 s and in the reminder of the experiment (right plot), the GNSS signals are significantly attenuated by the trees growing on both sides of the trail. During this part, the GNSS precision drops drastically, hovering around 1.6 m. From the mark 200 s, the sky view opens and the GNSS achieves its nominal precision again. On the other hand, the average precision of the total station reference is 7.0 mm as long as the robot lies in the field of view of the total stations.

These two experiments show that the proposed total station setup has an equivalent accuracy and precision compared to the RTK solution in an open area. However, our solution is significantly better in an obstructed environment such as a forest, where the GNSS signal is weaker because of the vegetation. Our setup can thus be used as a source of precise ground truth in both environment.

Reference

Vaidis, M., Giguère, P., Pomerleau, F., & Kubelka, V. (2021). Accurate outdoor ground truth based on total stations. 2021 18th Conference on Robots and Vision (CRV). https://arxiv.org/abs/2104.14396

Kilometer-scale autonomous navigation in subarctic forests: challenges and lessons learned

2021-11-30T00:00:00-05:00

Challenges inherent to autonomous wintertime navigation in forests include lack of reliable a Global Navigation Satellite System (GNSS) signal, low feature contrast, high illumination variations and changing environment. This type of off-road environment is an extreme case of situations au- tonomous cars could encounter in northern regions. Thus, it is important to understand the impact of this harsh environment on autonomous navigation systems. To this end, we present a field re- port analyzing teach-and-repeat navigation in a subarctic forest while subject to large variations of meteorological conditions. First, we describe the system, which relies on point cloud registration to localize a mobile robot through a boreal forest, while simultaneously building a map. We exper- imentally evaluate this system in over 18.8 km of autonomous navigation in the teach-and-repeat mode. Over 14 repeat runs, only four manual interventions were required, three of which were due to localization failure and another one caused by battery power outage. We show that dense vegetation perturbs the GNSS signal, rendering it unsuitable for navigation in forest trails. Further- more, we highlight the increased uncertainty related to localizing using point cloud registration in forest trails. We demonstrate that it is not snow precipitation, but snow accumulation, that affects our system’s ability to localize within the environment. Finally, we expose some challenges and lessons learned from our field campaign to support better experimental work in winter conditions

Link to the paper

Quick facts:

18.6 km of autonomous route repeating using a Clearpath Warthog
The description of our Weather-Invariant Lidar-based navigation (WILN) system, relying on a Robosense RS-32 lidar and a XSens MTi-10 IMU for autonomous navigation.
GNSS localization measurements using Emlid Reach-RS+ receivers.
Deployment in the Montmorency subarctic forest, the largest research forest in the world harsh winter weather.
We present a thorough analysis of the impact of the boreal forest biome and winter weather, inherent to northern countries on state-of-the-art autonomous navigation technologies.
We highlight the upcoming challenges in enabling multi-season autonomy in northern territories.
Our dataset is available to the community, you can access it here.

Video

Vincent Grondin

2021-07-27T00:00:00-04:00

Vincent Grondin is a Ph.D candidate supervised by Prof. Philippe Giguère at Université Laval, and he is a member of Norlab. He graduated with both a certificate in physics and a bachelor in electrical engineering from Université de Sherbrooke. During his bachelor, he was responsible for ensuring the control and simulation of a Hoverbike as a part of a higly-funded third year project. He also completed two internships at the Canadian Space Agency , where he developed a communication prototype with spatial recognition. Additionally, during an internship at the Speech and Audio Research Group Laboratory , he conducted research on the classification of speech and audio by using artificial intelligence (AI).

Currently, his research focuses on environment perception and simulation by employing AI in forests. Here is an example of his most recent project where AI algorithms perform automatic tree detection:

Education

M.Sc. en informatique ( Norlab ) - Université Laval (passage accéléré au doctorat), 2020
B.Ing. en génie électrique - Université de Sherbrooke, 2018
Certificat en physique - Université de Sherbrooke, 2014

Publications

Conference Articles

Baril, D., Grondin, V., Deschenes, S., Laconte, J., Vaidis, M., Kubelka, V., Gallant, A., Giguere, P., & Pomerleau, F. (2020). Evaluation of Skid-Steering Kinematic Models for Subarctic Environments. 2020 17th Conference on Computer and Robot Vision (CRV), 198–205. https://doi.org/10.1109/CRV50864.2020.00034

–>

Vincent Grondin

2021-07-27T00:00:00-04:00

Vincent Grondin est un étudiant au doctorat à l’Université Laval sous la supervision du Pr. Philippe Giguère et membre du laboratoire Norlab. Diplômé d’un certificat en physique et d’un baccalauréat en génie électrique à l’Université de Sherbrooke, il a occupé le poste de responsable en contrôle et asservissement d’un Hoverbike dans un projet majeur en ingénierie lors de sa dernière année au baccalauréat. Pendant ses études, il a réalisé deux stages à l’ Agence spatiale canadienne sur la recherche et développement d’un prototype de communication avec reconnaissance spatiale. Durant son stage au laboratoire Groupe de recherche sur la parole et l’audio , il a conduit une recherche sur la classification de la parole et l’audio à l’aide d’intelligence artificielle (IA).

Présentement, ses travaux de recherche se concentrent sur la simulation et la perception d’environnements par IA en milieux forestiers. Voici un exemple de détection automatique d’arbres:

Education

M.Sc. en informatique ( Norlab ) - Université Laval (passage accéléré au doctorat), 2020
B.Ing. en génie électrique - Université de Sherbrooke, 2018
Certificat en physique - Université de Sherbrooke, 2014

Publications

Conference Articles

Baril, D., Grondin, V., Deschenes, S., Laconte, J., Vaidis, M., Kubelka, V., Gallant, A., Giguere, P., & Pomerleau, F. (2020). Evaluation of Skid-Steering Kinematic Models for Subarctic Environments. 2020 17th Conference on Computer and Robot Vision (CRV), 198–205. https://doi.org/10.1109/CRV50864.2020.00034

–>

Norlab robots

2021-07-03T00:00:00-04:00

The list of our robots is growing, see for yourself! We like them rugged bacause they often go outdoors, collecting datasets, capturing lidar maps and driving autonomously around the campus and in the Montmorency forest. The black sheep of this family is the robotic arm, but even that is one day going to end up attached to the Warthog driving in some bushes 😃

Warthog

Large, all-terrain unmanned ground vehicle from Clearpath

It can handle tough environments with its rugged build, low ground pressure, and traction tires, which allow effortless mobility in soft terrain. In the winter season, we exchange the wheels for tracks that allow it to drive on all sorts of snow. It is the optimal platform for the forest environment the Norlab is all about.

Weight: 280 kg + payload (approx. 150 kg)
Size: 1.52 x 1.38 x 0.83 m
Max speed: 18 km/h
Battery: 12x lead-acid, 48 V, 2.5 h
Onboard computers: 2x automotive PC, 1x NVidia Xavier planned
Sensors: 2x 16-beam lidar, 1x 32-beam lidar, IMU, 2x RTK-GPS, 1x RGB camera

For our research, the warthog is equipped with tree RS-LIDARs from RoboSense, a Dalsa camera for collecting visual datasets, two GPS receivers for reference positioning and an IMU for precise attitude estimation. There are also two onboard PCs, one handling low-level drivers of the motor controllers and one dedicated for high-level tasks such as mapping and navigation. They run Ubuntu with ROS.

Husky

Skid-steered robotic development platform from Clearpath

The Husky’s rugged construction and high-torque drivetrain make it suitable for indoor and outdoor, moderately uneven terrains. Husky carries cameras, a LIDAR, and an IMU. It was the Norlab’s workhorse in the DARPA SubT Urban circuit and before that, the first explorer of the Montmorency forest.

Weight: 62 kg
Size: 0.99 x 0.67 x 0.39 m
Max speed: 3.6 km/h
Battery: 2x lead-acid, 24 V, 2 h
Onboard computers: 1x Dell 3070, 1x NVidia Xavier
Sensors: 1x 16-beam lidar, IMU, 6x RGB camera, 1x RGB-D camera

Its payload has been designed by Norlab's students for the DARPA SubT competition. The ability to explore the environment by an omnidirectional set of high-resolution cameras, by a lidar and by an additional RGB-D sensor makes this platform ideal for developing and testing machine-learning and SLAM algorithms. Although not perfect on obstacles and stairs, it can traverse rough terrain, and thus it is going to serve as a dataset-collecting platform of the lab.

Marmotte (HD2)

Tracked robotic development platform from SuperDroid Robots

The tracks of this platform are driven by powerful geared motors through heavy-duty chains. Its frame is ribbed and gusseted to make it rigid, and altogether it has a sturdy and solid chassis that will withstand rough treatment. Its size allows it to easily climb obstacles, ascend stairs, and drive over most terrains. In 2021, it is Norlab’s wild card in the DARPA SubT 🔥

Weight: 57 kg
Size: 1.00 x 0.65 x 0.50 m
Max speed: 4.0 km/h
Battery: 2x LiFePo4, 25.6 V, 2 h
Onboard computers: 1x Dell 3070, 1x NVidia Xavier
Sensors: 1x 16-beam lidar, IMU, 6x RGB camera, 1x RGB-D camera

This robot is Husky's young sister (Marmotte is she), equipped with the same sensor suite to accelerate software development for these two robots. Marmotte's advantage is the fact that it can climb stairs and get to places Husky cannot.

Vaul robot

Heavy unmanned ground vehicle from SuperDroid, customized by students to be a snowplow

The weight of this robot is perfect to plow snow on the sidewalks. The frame is made from aluminum, and four powerful geared motors allow the robot to carry a heavy payload. Students of the robotic club VAUL made a custom frame to carry the sensors and to protect the motors and batteries from the cold during the winter operations. With its plow, the robot can autonomously push around 40 kg of snow.

Weight: 160 kg
Size: 1.00 x 1.5 x 1.5 m
Max speed: 7.2 km/h
Battery: 4x Sealed Lead Acid Battery, 2x24V, 2h
Onboard computers: 1x Dell 3070, 1xNVidia Xavier AGX
Sensors: 1x 16-beam lidar, IMU, 3x RGB camera, 2xReach M+ RTK-GPS

With its diversified sensor suite, this snowplow can operate autonomously in diverse types of weather. The robot participated in its first Autonomous Snowplow Competition in Minneapolis in January 2020.

Universal Robots' UR10e

The UR10e is an extraordinarily versatile collaborative industrial robot from Universal Robots.

Our robot is equipped with a Robotiq 2-finger adaptive gripper. With a fixed or mounted-on-the-wrist camera, computer-vision-enabled applications make the UR10e perfect for pick-and-place and bin-picking applications.

Reach: 1300 mm
Payload: 12.5 kg
Footprint: diameter 190 mm
Repeatability: ±0.05 mm, with payload
Equipped with sensitive force & torque sensor
Our end-effector: Robotiq 2-finger adaptive gripper
Weight: 33.5 kg

The arm is equipped with 2D and 3D vision systems backed by powerful AI to fully unlock the potential of the robot to use state-of-the-art algorithms for pattern detection and object recognition, and design smart policies for manipulation tasks and autonomous grasping.

Lie theory cheatsheet

2021-06-25T00:00:00-04:00

Lie theory in robotics

In the last years, the Lie theory started to gain momentum and proved to be a powerful tool in the domain of state estimation. The theory was transposed in robotics terms in “State Estimation for Robotics” by Tim Barfoot. Also, a micro Lie theory was proposed by Sola et al..

Using the notations of Tim Barfoot’s book, we designed a cheatsheet synthesizing the main notions to remember when working with Lie groups.

Resources

Norlab in media!

2021-06-02T00:00:00-04:00

We are happy that Norlab is being noticed in media. See what they have written about us.

IEEE Spectrum:

IEEE Spectrum – Video Friday:

DARPA Subterranean Challenge:

Univesité Laval:

Norlab et ses partenaires se distinguent au DARPA Subterranean Challenge

Executive Gov:

DARPA Picks Eight Teams for Subterranean Challenge Robot Competition Finals

Center for Robotics and Autonomous Systems, Czech Technical University:

DARPA Sub-T

Robots & Automation News:

Darpa selects eight teams for $3.5 million prize competition

Geek Wire:

Robots and their masters take over nuclear plant for DARPA’s Subterranean Challenge

The Vidette:

Propulsion Québec – En Route! :

Laboratoire de robotique boréale (Norlab)

How to publish your preprints on Arxiv

2021-03-11T00:00:00-05:00

Before submitting a preprint to arXiv, make sure the publisher permits preprints. In general, they require that:

The authors disclose the existence of the preprint at submission.
Once an article is published, the preprint should link to the published version (typically via DOI or URL).

For IEEE, see IEEE publication Policies

For other publishers, see List of academic publishers by preprint policy

Submitting to arXiv

Register at arXiv.
Install arxiv-collector:
```
 pip install arxiv-collector	
```
In case of problems with the installation, you can download the standalone python file at this repository. You will also need a working installation of latexmk, see the documentation if it is not already installed in your system. You can also add the script directly to an Overleaf project and run it there for the same effect, please refer to the GitHub repository for instructions.
Execute arxiv-collector in the root directory of the Latex project:
```
 arxiv-collector root.tex
```
If using overleaf, arxiv-collector can be used online, see the documentation. In both cases, an archive arxiv.tar.gz will be created.
In your user page, click on START NEW SUBMISSION.
Review the user information and choose a license for your article. Be aware that some publishers may accept preprints on the condition that a specific license is used. In general, the CC BY: Creative Commons Attribution licence is fine.
Choose a primary classification for the submission. Most of the time, Computer Science - Robotics is a good choice.
Click Next and upload arxiv.tab.gz created beforehand. Click Continue:Process Files.
If the processing went well, you should see a Preview your paper button. Click on it and check that the paper did compile properly (pay attention to the figures and equations). If the processing yields error, which is quite frequent, see below for some troubleshooting.
On the next page, fill the metadata of the paper. Do not forget to fill the Comments field with additional information asked from the publisher. You can also specify that the paper has been submitted to the conference/journal.
Leave blank most of the last fields (i.e. Report Number, Journal Reference, DOI, ACM class, MSC class). You will be able to update these after the paper is published in the conference/journal.
Preview your submission and submit. It should take about a day for the paper to appear on arxiv.
Once the paper is on arxiv, an email will be sent to you with a password. Send it to your co-authors so that they can claim ownership of the paper as well.
Do not forget that once the paper is accepted in the conference/journal, you need to update the Comment field on arXiv with the publisher’s copyright.

Troubleshooting: arXiv Compilation

Most of the time, the compilation errors come from clashes in packages versions. An easy solution is to include the packages that may cause errors in the submission so that they are used in the compilation instead of the arXiv server’s packages.

Locate the package that cause errors in your system by adding the line \listfiles in your main latex file, and compile.
Open the log file root.log and search for the packages. You should find lines containing the paths to the packages. For instance:
```
 (/usr/share/texlive/texmf-dist/tex/latex/amsfonts/amsfonts.sty
 Package: amsfonts 2013/01/14 v3.01 Basic AMSFonts support
```
Copy the packages in the working directory and execute arxiv-collector root.tex. The packages are now included in the archive.
Re-upload the archive to arXiv and process the files.

List of known errors and solutions

Package biblatex-ieee Error: Failed to update citation style

Locate and include ieee.bbx and ieee.cbx in the working directory.

Four internships in robotics for the summer 2021 all taken!

2021-02-04T00:00:00-05:00

Norlab is opening four internships covering multiple fields of study for this summer. ~~Apply as soon as you can, we are closing the openings soon.~~ Here is the list of opportunities:

All the positions have been filled, but keep an eye on the Norlab website, there will be more in future!

Intern 1 - Not available anymore

computer science, software engineering, computer engineering

This internship is part of the SNOW (Self-driving Navigation Optimized for Winter) project, which aims at adapting autonomous navigation technologies to environments affected by winter. As part of his mandate, the student will develop an open library to predict the movements of a mobile robot. Several prototype algorithms have been analyzed in a Python Notebook, and must now be coded cleanly to allow real-time computation on a moving robot. The goal of the library is to allow researchers in the laboratory to do experiments faster. Ultimately, these experiments will allow mobile robots to adapt to changes in traction caused by snow cover during autonomous navigation.

The trainee will be supervised by a PhD student specializing in control, allowing the student to focus on software development while acquiring practical robotics skills. The student will also be expected to support the team’s efforts in robot integration and field deployments. The intern will have access to state-of-the-art equipment and a 500 kg robot. The library will be publicly hosted on the lab’s GitHub and will be integrated with ROS (Robot Operating System).

For contextual information, see the project page on SNOW.

Intern 2 - Not available anymore

computer science, software engineering, computer engineering

The laboratory is continuously developing several 3D mapping solutions for autonomous vehicles. Once the robot is on the move, it is relatively easy to locate it on a map. However, the laboratory has no software tools to relocate a robot if a map is reused.

The objective of the internship is to develop a software to relocate a robot in a given 3D map. The first step will be to relocalize the robot manually in a 3D graphical user interface (rviz). In a second step, the placement of the robot will have to be refined automatically using a recalibration tool used in the laboratory. Finally, an algorithm to find the optimal placement of the robot in the map automatically (Go-ICP) can be implemented. The student will be supervised by experts in 3D cartography and the registration algorithm. The code will be publicly hosted on the GitHub of the laboratory and will be integrated to ROS (Robot Operating System).

For contextual information, see the project page on libpointmatcher.

Intern 3 - Not available anymore

Electrical Engineering, Computer Engineering

The laboratory is preparing for a major U.S. deployment in September 2021. The goal of this deployment is to explore underground environments using a fleet of autonomous robots. To do so, the laboratory recently acquired new robots and commissioned the design of a custom power supply board for the sensors.

The objective of the internship is to complete the electronic integration of the robots and to participate in the robustness tests of the systems. The tracked robots should be able to climb stairs and withstand a certain level of shock. The student will be supervised by robotics experts with backgrounds in computer science, electricity and mechanics.

For contextual information, see the project page on SUBT.

Intern 4 - Not available anymore

computer science, software engineering, computer engineering

The laboratory is preparing for a major U.S. deployment in September 2021. The goal of this deployment is to explore underground environments using a fleet of autonomous robots. The laboratory is developing 3D mapping and localization solutions based on lidars, but the maps used are often too sparse to allow obstacle avoidance.

The objective of this internship is to combine the data from a lidar with the robot’s localization algorithms to produce a dense map within a close radius of the robot. In a first step, a dense map will be produced and displayed in a graphical user interface (rviz). Then, compression solutions will be tested to reduce the bandwidth of these maps during wireless transmission. Finally, obstacle avoidance algorithms will be studied. The student will be supervised by experts in 3D mapping and resetting algorithms. The code will be publicly hosted on the laboratory’s GitHub and will be integrated to ROS (Robot Operating System).

For contextual information, see the project page on SUBT.

Work environment

All internships will be held in the Norlab (Northern Robotics Laboratory). This research laboratory specializes in mobile and autonomous systems operating in northern or difficult conditions in general. We aim to study problems related to navigation algorithms in order to push the limits of what is currently possible to do with a mobile robot in real-life conditions. Currently, our efforts are directed towards localization algorithms based on laser sensors (lidar) and 3D reconstruction of environments.

The laboratory has two workshops in the Adrien-Pouliot building allowing interaction with research equipment according to current health standards. Unless otherwise specified, software development can be done remotely using existing data sets. Once the basis has been demonstrated, tests can be done on site with real research robots.

Compensation

There are two possible compensation options. The first option is for students aiming to continue at the Master’s level. If you have a rating greater than 3.8/4.3, we suggest that you apply for the NSERC Fellowship (https://www.nserc-crsng.gc.ca/students-etudiants/ug-pc/usra-brpc_fra.asp). Obtaining this scholarship will allow you to improve your record as a young researcher and normally comes with supplements from FRQNT. The laboratory also gives a $3,000 scholarship supplement for the internship. The application deadline is February 18, 2021.

If you are not selected for the award, salaries are set according to the collective agreement for research assistants and are approximately $15/hour. The work week is 35 hours/week. The duration of the internship may vary from 12 to 15 weeks.

Application

Send your choice of internship, your CV and your transcript to the laboratory director, François Pomerleau francois.pomerleau@ift.ulaval.ca .

Positions will be allocated at the beginning of March 2021.

Northern Robotics Laboratory

Lidar Scan Registration Robust to Extreme Motions

Contributions

Results in Images

Reference

Links

Accurate outdoor ground truth based on total stations

Contributions

Results in Images

Reference

Links

Kilometer-scale autonomous navigation in subarctic forests: challenges and lessons learned

Link to the paper

Quick facts:

Video

Vincent Grondin

Education

Publications

Conference Articles

Vincent Grondin

Education

Publications

Conference Articles

Norlab robots

Warthog

Large, all-terrain unmanned ground vehicle from Clearpath

Husky

Skid-steered robotic development platform from Clearpath

Marmotte (HD2)

Tracked robotic development platform from SuperDroid Robots

Vaul robot

Heavy unmanned ground vehicle from SuperDroid, customized by students to be a snowplow

Universal Robots' UR10e

The UR10e is an extraordinarily versatile collaborative industrial robot from Universal Robots.

Lie theory cheatsheet

Lie theory in robotics

Resources

Norlab in media!

How to publish your preprints on Arxiv

Submitting to arXiv

Troubleshooting: arXiv Compilation

List of known errors and solutions

Package biblatex-ieee Error: Failed to update citation style

Four internships in robotics for the summer 2021 all taken!

Intern 1 - Not available anymore

Intern 2 - Not available anymore

Intern 3 - Not available anymore

Intern 4 - Not available anymore

Work environment

Compensation

Application