Journal article Open Access

Uncertainty-Aware Imitation Learning using Kernelized Movement Primitives

João Silvério

JSON Export

  "files": [
      "links": {
        "self": ""
      "checksum": "md5:7762df9766d333922fc3b9d9f37040d1", 
      "bucket": "74e07bf2-022c-4743-9786-5440494647db", 
      "key": "Silverio_IROS19_2019.pdf", 
      "type": "pdf", 
      "size": 5783786
  "owners": [
  "doi": "10.5281/zenodo.3676791", 
  "stats": {
    "version_unique_downloads": 49.0, 
    "unique_views": 25.0, 
    "views": 31.0, 
    "version_views": 51.0, 
    "unique_downloads": 30.0, 
    "version_unique_views": 30.0, 
    "volume": 196648724.0, 
    "version_downloads": 65.0, 
    "downloads": 34.0, 
    "version_volume": 378090120.0
  "links": {
    "doi": "", 
    "conceptdoi": "", 
    "bucket": "", 
    "conceptbadge": "", 
    "html": "", 
    "latest_html": "", 
    "badge": "", 
    "latest": ""
  "conceptdoi": "10.5281/zenodo.3676692", 
  "created": "2020-02-20T13:05:10.738068+00:00", 
  "updated": "2020-02-21T08:25:26.465637+00:00", 
  "conceptrecid": "3676692", 
  "revision": 5, 
  "id": 3676791, 
  "metadata": {
    "access_right_category": "success", 
    "doi": "10.5281/zenodo.3676791", 
    "description": "<p>During the past few years, probabilistic approaches to imitation learning have earned a relevant place in the robotics literature. One of their most prominent features is that, in addition to extracting a mean trajectory from task demonstrations, they provide a variance estimation. The intuitive meaning of this variance, however, changes across different techniques, indicating either variability or uncertainty. In this paper we leverage kernelized movement primitives (KMP) to provide a new perspective on imitation learning by predicting variability, correlations and uncertainty using a single model. This rich set of information is used in combination with the fusion of optimal controllers to learn robot actions from data, with two main advantages: i) robots become safe when uncertain about their actions and ii) they are able to leverage partial demonstrations, given as elementary sub-tasks, to optimally perform a higher level, more complex task. We showcase our approach in a painting task, where a human user and a KUKA robot collaborate to paint a wooden board. The task is divided into two sub-tasks and we show that the robot becomes compliant (hence safe) outside the training regions and executes the two sub-tasks with optimal gains otherwise</p>", 
    "language": "eng", 
    "title": "Uncertainty-Aware Imitation Learning using Kernelized Movement Primitives", 
    "license": {
      "id": "CC-BY-4.0"
    "relations": {
      "version": [
          "count": 3, 
          "index": 2, 
          "parent": {
            "pid_type": "recid", 
            "pid_value": "3676692"
          "is_last": true, 
          "last_child": {
            "pid_type": "recid", 
            "pid_value": "3676791"
    "communities": [
        "id": "collaborate_project"
    "publication_date": "2019-09-30", 
    "creators": [
        "affiliation": "Idiap Research Institute, Martigny, Switzerland", 
        "name": "Jo\u00e3o Silv\u00e9rio"
    "access_right": "open", 
    "resource_type": {
      "subtype": "article", 
      "type": "publication", 
      "title": "Journal article"
    "related_identifiers": [
        "scheme": "doi", 
        "identifier": "10.5281/zenodo.3676692", 
        "relation": "isVersionOf"
All versions This version
Views 5131
Downloads 6534
Data volume 378.1 MB196.6 MB
Unique views 3025
Unique downloads 4930


Cite as