"Better Algorithms – Cheaper Precision Rockets"
For a rocket to hit its intended target, it must know where it is coming from and where it is going. In order to develop precision rockets that would also carry a competitive price tag, you must be creative. At IMI Systems, Dr. Agnes Cohen deals with the challenge using computer simulation
Ami Rojkes Dombe
| 21/12/2017
Photo: IMI Systems
One of the challenges currently facing most Israeli defense industries is the question of how to remain a leader among leaders at a competitive price. This challenge has not skipped IMI Systems, which is involved – among other fields – in the field of precision rockets and mortar bombs. At IMI Systems, they decided to deal with this challenge by using computer simulation.
For the past two decades, IMI has had a GNC (Guidance, Navigation & Control) project department that deals with simulation, control and algorithms. Over the last few years, this 15-employee department has been responsible for the development of algorithms for all of IMI Systems' divisions and units. The department is a part of IMI Systems' Fire Systems Division. Heading the department is Dr. Agnes Cohen, who came to IMI from the Technion – Israel Institute of Technology more than twenty years ago.
"The department develops algorithms in three primary fields – control, guidance and navigation," explains Dr. Cohen. "Rockets should be accurate and you need a simulation with models of physical processes that would help us achieve that goal at a competitive cost. The objective of the algorithms is to get the rocket or shell to the target subject to known and unknown constraints."
At IMI Systems they explain that for a rocket to hit its intended target, it must know where it is coming from, where it is going, its current position and its orientation. While in the past the Company had to carry out extremely costly physical tests primarily, the progress made in the field of computer capabilities has changed this approach, and today most of their tests are carried out using computer simulation. "The entire firing process is programmed into the computer in the form of models," explains Dr. Cohen. "We program models of all rocket elements, the flight computer, the sensors, the engine, the servo motors and any other element that may affect the operation of the rocket. Subsequently, we start testing the operation of the rocket under different conditions.
"What we are dealing with is a long, slender rocket to which guidance was added, which functions within a specific performance envelope. It can maneuver in a certain way, structurally and aerodynamically. One of our objectives at the department is to keep it within a suitable performance envelope. If you issue a guidance command but do not have effective control, the rocket might disintegrate. The control element enforces the command received from the command and control system and ensures the circuit closes during operation.
"Every guided rocket contains an INS and GPS unit and a navigation algorithm, but that is not always enough. We must know the physical characteristics and variables of the rocket, including acceleration rates, angular rates and other data. There are accelerometers, rate meters and a servo motor – an electrical motor designed to receive instructions and close the circuit while the physical data change constantly during flight. You issue a command and until the circuit closes a physical process takes place which includes an element of uncertainty. At this point, the control algorithms developed using the simulation come into play. This is the difference between a rocket capable of providing a guaranteed CEP and a rocket that deviates from that parameter.
"One should bear in mind the fact that all of the sensors come with noise and errors. We must model it very precisely and develop algorithms that would correct the rocket's stability and uncertainty gaps. We develop the control algorithm around all that. Additionally, there is an algorithm for guidance and analyzing the mission energetics. If you have a vehicle you can launch to a range of X kilometers and you decide to launch it to a range of 2X kilometers, the command and control system will tell you it is outside of the energetic range. Our algorithms will provide the alert.
Close Cooperation with the Field
One of Dr. Cohen's responsibilities is supporting IMI's marketing representatives when they are required to answer questions presented by clients. For example, when a client wants to know what would happen in a scenario where the rocket had lost the external guidance capability – the guaranteed CEP to which IMI will commit in such a situation. "We do that using the simulation," explains Dr. Cohen. "We also use faster feasibility tests that do not require extensive calculations, only to obtain a basic idea as to the client's demands.
"Our advantage opposite the competition stems from our cooperation with the development departments throughout IMI and with the field elements. For example, we cooperate with the engine development department. If the simulation indicates that a rocket needs more propulsion or other capabilities, the information will be submitted to engine development and they will adapt the engine to the requirements of the rocket.
"As part of the algorithm development activity, we also deal with neural networks and are well-versed in the academic research activity in this particular field. At the same time, our development is intended for weapon systems exclusively, so the development activity is based mostly on methods tested and proven within IMI Systems."
At IMI Systems they explain that in the field, clients sometimes want to use rockets or shells with their own existing infrastructures. In such cases, the systems in question sometime consist of relatively inexpensive hardware where the meters do not provide accurate information. "One of the challenges in such cases is to reevaluate data without accurate feedback from the sensors," explains Dr. Cohen. "We need to reevaluate on the basis of correlation of multiple data from different sensors. We know how to deal with inexpensive meters, namely – inexpensive hardware that is less accurate. The data are not as good as they should be and we must squeeze the algorithms to the limit in order to achieve the same results. It is a trade-off between quality electronics and algorithms. It is a challenge for the entire industry.
"The higher the algorithm quality, the cheaper the rocket will be. We at IMI know how to deliver a rocket to the target using cellular phone meters if so required. We work closely with the system engineering elements and deal with challenges from the field".
Another challenge emerges whenever a new rocket or artillery shell is introduced. In these cases, new algorithms should be designed based on the knowledge available to IMI Systems. "We can do it within three months – to develop a system you can start working with, a new simulation for the trials," says Dr. Agnes Cohen.
"The department also deals with minimizing the collateral damage of the rocket using algorithms. You can issue a command to the rocket not to explode if it had digressed from its predetermined operational envelope. There are definitions for safety margins. The mechanism knows where you are located relative to your performance envelope. If you digress – the command and control system will be alerted.
"Prior to an actual physical trial we conduct thousands of tests and we also use hybrid simulation. The hybrid simulator is a special room that enables us to connect the flight computer to the rocket's sensors and servo motors. In this way, we can simulate the actual data according to the client's actual meters. This will make the simulation almost identical to a physical trial.
For a rocket to hit its intended target, it must know where it is coming from and where it is going. In order to develop precision rockets that would also carry a competitive price tag, you must be creative. At IMI Systems, Dr. Agnes Cohen deals with the challenge using computer simulation
Photo: IMI Systems
One of the challenges currently facing most Israeli defense industries is the question of how to remain a leader among leaders at a competitive price. This challenge has not skipped IMI Systems, which is involved – among other fields – in the field of precision rockets and mortar bombs. At IMI Systems, they decided to deal with this challenge by using computer simulation.
For the past two decades, IMI has had a GNC (Guidance, Navigation & Control) project department that deals with simulation, control and algorithms. Over the last few years, this 15-employee department has been responsible for the development of algorithms for all of IMI Systems' divisions and units. The department is a part of IMI Systems' Fire Systems Division. Heading the department is Dr. Agnes Cohen, who came to IMI from the Technion – Israel Institute of Technology more than twenty years ago.
"The department develops algorithms in three primary fields – control, guidance and navigation," explains Dr. Cohen. "Rockets should be accurate and you need a simulation with models of physical processes that would help us achieve that goal at a competitive cost. The objective of the algorithms is to get the rocket or shell to the target subject to known and unknown constraints."
At IMI Systems they explain that for a rocket to hit its intended target, it must know where it is coming from, where it is going, its current position and its orientation. While in the past the Company had to carry out extremely costly physical tests primarily, the progress made in the field of computer capabilities has changed this approach, and today most of their tests are carried out using computer simulation. "The entire firing process is programmed into the computer in the form of models," explains Dr. Cohen. "We program models of all rocket elements, the flight computer, the sensors, the engine, the servo motors and any other element that may affect the operation of the rocket. Subsequently, we start testing the operation of the rocket under different conditions.
"What we are dealing with is a long, slender rocket to which guidance was added, which functions within a specific performance envelope. It can maneuver in a certain way, structurally and aerodynamically. One of our objectives at the department is to keep it within a suitable performance envelope. If you issue a guidance command but do not have effective control, the rocket might disintegrate. The control element enforces the command received from the command and control system and ensures the circuit closes during operation.
"Every guided rocket contains an INS and GPS unit and a navigation algorithm, but that is not always enough. We must know the physical characteristics and variables of the rocket, including acceleration rates, angular rates and other data. There are accelerometers, rate meters and a servo motor – an electrical motor designed to receive instructions and close the circuit while the physical data change constantly during flight. You issue a command and until the circuit closes a physical process takes place which includes an element of uncertainty. At this point, the control algorithms developed using the simulation come into play. This is the difference between a rocket capable of providing a guaranteed CEP and a rocket that deviates from that parameter.
"One should bear in mind the fact that all of the sensors come with noise and errors. We must model it very precisely and develop algorithms that would correct the rocket's stability and uncertainty gaps. We develop the control algorithm around all that. Additionally, there is an algorithm for guidance and analyzing the mission energetics. If you have a vehicle you can launch to a range of X kilometers and you decide to launch it to a range of 2X kilometers, the command and control system will tell you it is outside of the energetic range. Our algorithms will provide the alert.
Close Cooperation with the Field
One of Dr. Cohen's responsibilities is supporting IMI's marketing representatives when they are required to answer questions presented by clients. For example, when a client wants to know what would happen in a scenario where the rocket had lost the external guidance capability – the guaranteed CEP to which IMI will commit in such a situation. "We do that using the simulation," explains Dr. Cohen. "We also use faster feasibility tests that do not require extensive calculations, only to obtain a basic idea as to the client's demands.
"Our advantage opposite the competition stems from our cooperation with the development departments throughout IMI and with the field elements. For example, we cooperate with the engine development department. If the simulation indicates that a rocket needs more propulsion or other capabilities, the information will be submitted to engine development and they will adapt the engine to the requirements of the rocket.
"As part of the algorithm development activity, we also deal with neural networks and are well-versed in the academic research activity in this particular field. At the same time, our development is intended for weapon systems exclusively, so the development activity is based mostly on methods tested and proven within IMI Systems."
At IMI Systems they explain that in the field, clients sometimes want to use rockets or shells with their own existing infrastructures. In such cases, the systems in question sometime consist of relatively inexpensive hardware where the meters do not provide accurate information. "One of the challenges in such cases is to reevaluate data without accurate feedback from the sensors," explains Dr. Cohen. "We need to reevaluate on the basis of correlation of multiple data from different sensors. We know how to deal with inexpensive meters, namely – inexpensive hardware that is less accurate. The data are not as good as they should be and we must squeeze the algorithms to the limit in order to achieve the same results. It is a trade-off between quality electronics and algorithms. It is a challenge for the entire industry.
"The higher the algorithm quality, the cheaper the rocket will be. We at IMI know how to deliver a rocket to the target using cellular phone meters if so required. We work closely with the system engineering elements and deal with challenges from the field".
Another challenge emerges whenever a new rocket or artillery shell is introduced. In these cases, new algorithms should be designed based on the knowledge available to IMI Systems. "We can do it within three months – to develop a system you can start working with, a new simulation for the trials," says Dr. Agnes Cohen.
"The department also deals with minimizing the collateral damage of the rocket using algorithms. You can issue a command to the rocket not to explode if it had digressed from its predetermined operational envelope. There are definitions for safety margins. The mechanism knows where you are located relative to your performance envelope. If you digress – the command and control system will be alerted.
"Prior to an actual physical trial we conduct thousands of tests and we also use hybrid simulation. The hybrid simulator is a special room that enables us to connect the flight computer to the rocket's sensors and servo motors. In this way, we can simulate the actual data according to the client's actual meters. This will make the simulation almost identical to a physical trial.
