{
  "id": 110537,
  "title": "Why does the same sample have different ego pose tokens for each sensor?",
  "url": "/competitions/3d-object-detection-for-autonomous-vehicles/discussion/110537",
  "author_name": "Yi23333",
  "post_date": "2019-09-28T20:46:47.537000",
  "votes": 3,
  "comment_count": 3,
  "views": 0,
  "content": "<p>We know that for one \"sample\", there are 8 or 10 sensors at one same \"timestamp\". They are supposed to be on the same car, and therefore have one \"ego pose\". However in the data, each sensor has its own ego pose token. The rotation and translation values are close, but different. </p>\n\n<p>For example, the following two ego poses are from one sample: <br>\n{'rotation': [0.8077788684109727,\n  0.005242205394068225,\n  0.0002022063339964093,\n  0.589462278813406],\n 'translation': [1101.9188041855564, 2027.1876964550627, -24.793249010626884],\n 'token': '5cfdefb521ae5cf47c99db2044615379b20eb0b059412e69d5e2221ef28fb1da',\n 'timestamp': 1552002939100959.2} \nand <br>\n{'rotation': [0.8079349267472453,\n  0.0055296651765656575,\n  0.0001167004720639972,\n  0.5892457580042119],\n 'translation': [1101.823331478092, 2026.8231066845985, -24.77530010914645],\n 'token': 'dd442f2f99e8c4ecb6a8f5999f469e59b50e66e9baf229ff045c7b1bfa6e173f',\n 'timestamp': 1552002939100959.2}</p>",
  "messages": [
    {
      "id": 636120,
      "postDate": "2019-09-28T20:46:47.537Z",
      "content": "<p>We know that for one \"sample\", there are 8 or 10 sensors at one same \"timestamp\". They are supposed to be on the same car, and therefore have one \"ego pose\". However in the data, each sensor has its own ego pose token. The rotation and translation values are close, but different. </p>\n\n<p>For example, the following two ego poses are from one sample: <br>\n{'rotation': [0.8077788684109727,\n  0.005242205394068225,\n  0.0002022063339964093,\n  0.589462278813406],\n 'translation': [1101.9188041855564, 2027.1876964550627, -24.793249010626884],\n 'token': '5cfdefb521ae5cf47c99db2044615379b20eb0b059412e69d5e2221ef28fb1da',\n 'timestamp': 1552002939100959.2} \nand <br>\n{'rotation': [0.8079349267472453,\n  0.0055296651765656575,\n  0.0001167004720639972,\n  0.5892457580042119],\n 'translation': [1101.823331478092, 2026.8231066845985, -24.77530010914645],\n 'token': 'dd442f2f99e8c4ecb6a8f5999f469e59b50e66e9baf229ff045c7b1bfa6e173f',\n 'timestamp': 1552002939100959.2}</p>",
      "rawMarkdown": "We know that for one \"sample\", there are 8 or 10 sensors at one same \"timestamp\". They are supposed to be on the same car, and therefore have one \"ego pose\". However in the data, each sensor has its own ego pose token. The rotation and translation values are close, but different. \n\nFor example, the following two ego poses are from one sample:  \n{'rotation': [0.8077788684109727,\n  0.005242205394068225,\n  0.0002022063339964093,\n  0.589462278813406],\n 'translation': [1101.9188041855564, 2027.1876964550627, -24.793249010626884],\n 'token': '5cfdefb521ae5cf47c99db2044615379b20eb0b059412e69d5e2221ef28fb1da',\n 'timestamp': 1552002939100959.2} \nand  \n{'rotation': [0.8079349267472453,\n  0.0055296651765656575,\n  0.0001167004720639972,\n  0.5892457580042119],\n 'translation': [1101.823331478092, 2026.8231066845985, -24.77530010914645],\n 'token': 'dd442f2f99e8c4ecb6a8f5999f469e59b50e66e9baf229ff045c7b1bfa6e173f',\n 'timestamp': 1552002939100959.2}",
      "votes": 3
    },
    {
      "id": 636129,
      "postDate": "2019-09-28T20:57:59.663Z",
      "content": "<p>Good question, ego pose is the position of the vehicle in global coordinate system, it should be same for all the sensors on a given ego vehicle. </p>",
      "rawMarkdown": "Good question, ego pose is the position of the vehicle in global coordinate system, it should be same for all the sensors on a given ego vehicle. ",
      "votes": 1
    },
    {
      "id": 636368,
      "postDate": "2019-09-29T11:47:43.393Z",
      "content": "<p>Maybe this is because the vehicle is moving and data is captured from sensors at different times to be maximally in sync with the lidar? So ego pose also changes.</p>",
      "rawMarkdown": "Maybe this is because the vehicle is moving and data is captured from sensors at different times to be maximally in sync with the lidar? So ego pose also changes.",
      "votes": 2
    },
    {
      "id": 636406,
      "postDate": "2019-09-29T13:28:54.877Z",
      "rawMarkdown": "",
      "votes": 3,
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 636129,
      "author_name": "Rishabh Agrahari",
      "author_url": "",
      "post_date": "2019-09-28T20:57:59.663000",
      "content": "<p>Good question, ego pose is the position of the vehicle in global coordinate system, it should be same for all the sensors on a given ego vehicle. </p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 636368,
      "author_name": "Konstantin Lopukhin",
      "author_url": "",
      "post_date": "2019-09-29T11:47:43.393000",
      "content": "<p>Maybe this is because the vehicle is moving and data is captured from sensors at different times to be maximally in sync with the lidar? So ego pose also changes.</p>",
      "votes": 2,
      "replies": []
    },
    {
      "id": 636406,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-09-29T13:28:54.877000",
      "content": "",
      "votes": 3,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "636120": "We know that for one \"sample\", there are 8 or 10 sensors at one same \"timestamp\". They are supposed to be on the same car, and therefore have one \"ego pose\". However in the data, each sensor has its own ego pose token. The rotation and translation values are close, but different. \n\nFor example, the following two ego poses are from one sample:  \n{'rotation': [0.8077788684109727,\n  0.005242205394068225,\n  0.0002022063339964093,\n  0.589462278813406],\n 'translation': [1101.9188041855564, 2027.1876964550627, -24.793249010626884],\n 'token': '5cfdefb521ae5cf47c99db2044615379b20eb0b059412e69d5e2221ef28fb1da',\n 'timestamp': 1552002939100959.2} \nand  \n{'rotation': [0.8079349267472453,\n  0.0055296651765656575,\n  0.0001167004720639972,\n  0.5892457580042119],\n 'translation': [1101.823331478092, 2026.8231066845985, -24.77530010914645],\n 'token': 'dd442f2f99e8c4ecb6a8f5999f469e59b50e66e9baf229ff045c7b1bfa6e173f',\n 'timestamp': 1552002939100959.2}",
    "636129": "Good question, ego pose is the position of the vehicle in global coordinate system, it should be same for all the sensors on a given ego vehicle. ",
    "636368": "Maybe this is because the vehicle is moving and data is captured from sensors at different times to be maximally in sync with the lidar? So ego pose also changes.",
    "636406": ""
  }
}