Removable automotive LIDAR data collection pod
A vehicle agnostic removable pod can be mounted on a vehicle using one or more legs of a pod mount. The removable pod can collect and time stamp a variety of environmental data as well as vehicle data. For example, environmental data can be collected using a sensor suite which can include an IMU, 3D positioning sensor, one or more cameras, and/or a LIDAR unit. As another example, vehicle data can be collected via a CAN bus attached to the vehicle. Environmental data and/or vehicle data can be time stamped and transmitted to a remote server for further processing by a computing device.
1. A method of collecting diverse driving data, the method comprising:
for each of a plurality of removable pods each respectively attached to a unique vehicle in an environment:
collecting respective environmental data of the environment using one or more sensors of a respective removable pod,
recording the respective environmental data collected using the respective removable pod and corresponding vehicle data that corresponds to the respective environmental data,
determining time stamps for each instance of the respective environmental data and time stamps for each instance of the corresponding vehicle data using a respective computing device connected to the respective removable pod,
generating a respective aggregation of the time stamped instances of the respective environmental data and the time stamped instances of the corresponding vehicle data using the respective computing device, and
transmitting, to a remote server, the respective aggregation of the time stamped instances of the respective environmental data and the time stamped instances of the corresponding vehicle data;
aggregating, at the remote server, the time stamped instances of vehicle data and environmental data from the plurality of removable pods,
wherein each of the unique vehicles is attached to at least one removable pod;
generating training data for a machine learning model based on the time stamped instances of vehicle data and environmental data from the plurality of removable pods; and
training, based on the training data, the machine learning model, to generate control signals for an autonomous vehicle.
2. The method of claim 1 , wherein recording the respective environmental data and the corresponding vehicle data for the respective removable pod is performed using the respective computing device that is connected to the respective removable pod.
3. The method of claim 1 , wherein recording the respective environmental data and the corresponding vehicle data for the respective removable pod occurs on a separate storage device operatively connected to a vehicle cabin removable pod hardware of the respective removable pod.
4. The method of claim 1 , wherein the unique vehicles each being a non-autonomous vehicle that are collecting data in the environment with the autonomous vehicle or an additional autonomous vehicle.
5. The method of claim 4 , further comprising:
automatically labeling data collected from the autonomous vehicle or the additional autonomous vehicle using data collected from at least one removable pod, out of the plurality of removable pods, at a known location and time.
6. The method of claim 1 , wherein determining the time stamps for each instance of the respective environmental data and the time stamps for each instance of the corresponding vehicle data comprises:
determining computing device time stamps for at least one instance of the environmental or vehicle data, and
determining the time stamps for the at least one instance of the environmental or vehicle data based on having the computing device time stamps synched with time stamps from one or more sensors, or a printed circuit board (PCB), of the respective removable pod.
7. The method of claim 1 , wherein the respective aggregation of the time stamped instances of environmental data and the time stamped instances of corresponding vehicle data for the respective removable pod is generated with an external data source.
8. The method of claim 1 , wherein the unique vehicles include at least two vehicles of different vehicle types.
9. The method of claim 1 , wherein the unique vehicles include vehicles having variations and within a vehicle type.
10. The method of claim 1 , wherein the plurality of removable pods are vehicle-agnostic.
11. The method of claim 1 , wherein more than one removable pod is attached to a vehicle, of the unique vehicles.
12. The method of claim 1 , wherein the plurality of removable pods are each mounted at a valid position of a corresponding vehicle to allow sensors of the plurality of removable pods to collect data from a known predefined perspective.
13. The method of claim 1 , wherein the unique vehicles include the autonomous vehicle or an additional autonomous vehicle.
14. A method of collecting diverse driving data, the method comprising:
receiving, at a remote server and for each of a plurality of removable pods that are each respectively attached to a unique vehicle in an environment, a respective aggregation of time stamped instances of environmental data and vehicle data,
wherein each of the vehicles is attached to at least one removable pod;
aggregating, at the remote server, the time stamped instances of vehicle data and environmental data from the plurality of removable pods,
wherein prior to receiving the aggregations for the plurality of removable pods, and for a respective removable pod of the plurality of removable pods,
respective environmental data of the environment is collected using one or more sensors of the respective removable pod,
the respective environmental data and corresponding vehicle data is recorded,
the time stamps for each instance of the respective environmental data and time stamps for each instance of the corresponding vehicle data are determined using a respective computing device connected to the respective removable pod, and
a respective aggregation of the time stamped instances of respective environmental data and the time stamped instances of corresponding vehicle data is generated using the respective computing device;
generating training data for a machine learning model based on the time stamped instances of vehicle data and environmental data from the plurality of removable pods; and
training, based on the training data, the machine learning model, to generate control signals for an autonomous vehicle.
15. The method of claim 14 , wherein the respective environmental data and corresponding vehicle data for the respective removable pod is recorded using the respective computing device that is connected to the respective removable pod.
16. The method of claim 14 , wherein the respective environmental data and the corresponding vehicle data for the respective removable pod is recorded on a separate storage device operatively connected to a vehicle cabin removable pod hardware of the respective removable pod.
17. The method of claim 14 , wherein the multiple vehicles each being a non-autonomous vehicle that are collecting data in the environment with the autonomous vehicle or an additional autonomous vehicle.
18. The method of claim 17 , further comprising:
automatically labeling data collected from the autonomous vehicle or the additional autonomous vehicle using data collected from at least one removable pod, out of the plurality of removable pods, at a known location and time.