Motion Classification of Objects using Accelerometer and Gyroscope Readings
Bibliographic record
Abstract
It can be challenging to properly analyze data from moving objects when there is limited information about the conditions under which the data was collected. The focus of this research is to use accelerometer and gyroscope data as the input to machine learning models to accurately classify data points based on the movement of the object within its environment. Three machine learning models were implemented and tested: a Hidden Markov Model (HMM), a Random Forest Model (RF), and a Multilayer Perceptron Model (MLP). The majority of the data collected via a simulated environment in Gazebo was used to train all three models. The simulated experiments involved a Jackal Robot from Clearpath Robotics driving on three terrains — flat ground and up and down a ramp with a 7-degree slope — that match the classification states. The simulation data was manually classified based on the delta values between orientation data. The models were tested using the remaining 20% of simulated data, data from a physical Jackal, and data from haul trucks working in an open-pit mine. The HMM is based on unsupervised learning which meant that the pre-classified datapoints were not useful in training the model. This model was ineffective at classifying the datapoints because the probabilities of the classification changing from one datapoint to the next were situationally skewed, so it was highly improbable that the state changed between datapoints. Due to the inability of this model to classify the simulated data, no further testing was completed. The RF model was able to classify the simulated data with 99.78% accuracy and the MLP model classified the simulated data with 99.73% accuracy. The RF model and the MLP model require noisier training data to accurately classify real-world data. Further testing is currently being conducted using data from a physical Jackal and haul trucks.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".