Design and development of autonomous grader and algorithm

Unmanned machines are used for leveling works such as removal of nuclear residues, road leveling applications for military roads, or for leveling tasks that requires unmanned intervention where the risk of occupational hazard is high. Earthworks are unique when the soil type and terrain are considered. As a result, each earthwork task is unique. As autonomous vehicles become more widespread in transportation, different solutions are being proposed for construction machinery. Unmanned applications are seen more often in applications such as open mine or leveling excavations which can be programmed. In this thesis, an algorithm is developed for the autonomous leveling of a terrain whose elevation can be entered as input data. Terrain is excavated to the desired level by the autonomous machine with a blade width of 1 m. With the developed algorithm, the vehicle finds the closest local maximum elevation and starts grading to the boundary point where it started. This process is repeated until the strip is excavated to the desired elevation. In order to test the algorithm, terrain models with different heights were created with random numbers and divided into small squares in grid form. Initial values are assigned for chassis weight, number of batteries and engine power etc. in the beginning and inadequate components are improved. Excavation thickness and blade capacity of the vehicle were kept below a certain limit. The height of the excavated area was updated and the terrain is leveled. The time of the shoveling process, the excavated volume and the drained volume were calculated. Forces and resistances acting on the construction machine caused by the excavation, inclination, rolling, friction, ground and drag resistances that occurred were calculated dynamically. The required power is calculated based on the calculated resistive force and design speed. It was decided whether or not it was capable of doing the leveling task by comparing it with the engine power assigned in the beginning. Depending on the power and calculated time, the energy requirement thus, the number of required batteries was calculated. Initially assigned number of batteries and the engine powers are compared by the results of analysis. In case the engine and battery power is not sufficient, insufficient components are strengthened and the simulation of leveling continued until all components are adequate. Three different leveling simulations and nine different soil-machine interaction modeling simulations were performed. Three different velocities are assigned for the autonomous grade for three different soil types. As the stiffness of the land increases, the required motor power increases. However, the number of batteries decreases as the vehicle weight decreases. The cost of the construction machine which achieves the design objectives is calculated by simulation. This study can be used as a reference for a manufacturer to estimate the cost of the construction machine. The leveling process was carried out by the unmanned machine with certain limitations without human intervention. The autonomous shovel tool study has been tried to find the best alternative depending on the power and battery options of the iteration process. Thanks to this study, the leveling practices requiring unmanned intervention will be solved without increasing the safety risk. The vehicle's engine capacity, battery power and chassis will be highly efficient. As the engine power and number of batteries will be determined to meet the design criteria, the production cost of the machine will be reduced. In the simulation operation, the most suitable engine power, number of batteries and chassis weight were obtained without any human intervention.